Medidas_de_Performance_Avaliacao_Modelos.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Medidas_de_Performance_Avaliacao_Modelos.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/clemencia/ML4PPGF_UERJ/blob/master/Medidas_de_Performance_Avaliacao_Modelos.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vEl1j0yhe5Im",
        "colab_type": "text"
      },
      "source": [
        "# ROC: Avaliação de Classificadores Binários\n",
        "\n",
        "Em problemas de classificação binária (distinguir sinal de ruído)  a medida de performance frequentemente é a **acurácia**, porém as vezes uma alta acurácia vem a um custo  de **ineficiência**. Por isso em casos de classificadores como BDT vimos as curvas PRC que indicavam como muda a **precisão** (ou **pureza**)  e o ***recall*** ou **eficiência** ao mudarmos o critério de corte.\n",
        "\n",
        "---\n",
        "\n",
        "### A curva ROC\n",
        "A sigla ROC vem de ***receiving operator characteristic***, pois foi ideado originalmente para detectar objetos em radares, ou também ***relative operator characteristic*** pois relaciona duas caraterísticas do \"operador\", a precisão e a eficiência.\n",
        "\n",
        "Primeiro devemos esclarecer as definições dos conceitos acima. Vamos ter em mente a *matriz confusão*:\n",
        "\n",
        "\n",
        "|*Resposta*| Sinal | Ruido | ***Total*** |\n",
        "|----------| ------|------ |-------------|\n",
        "|*Negativo*|  <font color=red >$F_N$ </font>| <font color=green> $V_N$</font> |         $N$ |\n",
        "|*Positivo*| <font color=green >$V_P$</font> | <font color=red> $F_P$</font>  |        $P$ |\n",
        "|**Total** | $$S$$   |  $$B $$  |  $$T = N+P = S+B$$  |\n",
        "\n",
        "\n",
        "Onde a soma das linhas:\n",
        " $$ P = V_P + F_P  \\;\\;\\; \\mathrm{e} \\;\\;\\; N = F_N + V_N $$\n",
        "e a soma das colunas:\n",
        " $$ S = V_P + F_N  \\;\\;\\; \\mathrm{e} \\;\\;\\; B = V_N + F_P $$\n",
        "\n",
        "---\n",
        "\n",
        "### Definições : \n",
        "   * **Acurácia**: fração das instâncias classificadas corretamente (seja como sinal ou background)\n",
        "$$ a = \\dfrac{V_P + V_N}{T} $$\n",
        "   * **Precisão**: fração das instâncias classificadas como positivas que são sinal verdadeiro (i.e. estão corretas). Também chamada de *pureza*.\n",
        "   $$ p = \\dfrac{V_P}{P}$$\n",
        "   * **Eficiência**: fração do sinal que foi corretamente classificado de forma positiva. Também chamado de *recall* ou **TPR** (*true positive rate*).\n",
        "   $$ \\epsilon = \\dfrac{V_P}{S}$$\n",
        "   * **Especificidade**: fração de background corretamente classificado de forma negativa, também conhecido de  **TNR** (*true negative rate*) e seria equivalente à eficiência em identificar o ruído (aka *background rejection*)\n",
        "   $$ \\epsilon_{B} = \\dfrac{V_N}{B} $$  \n",
        "   * ***False Positive Rate***: (**FPR**) é a fração das instâncias de background classificadas incorretamente de forma positiva.\n",
        "   $$ f_{FP} = \\dfrac{F_P}{B} = 1-\\epsilon_B$$\n",
        "   * ***False Negative Rate***: (**FNR**) é a fração das instâncias de sinal classificadas incorretamente de forma negativa, i.e. a *ineficiência*.\n",
        "   $$ f_{FN} = \\dfrac{F_N}{S} = 1 - \\epsilon$$ \n",
        "\n",
        "---\n",
        "\n",
        "Nas curvas ROC geralmente  é graficado no eixo $y$ a eficiência (aka recall ou TPR) que queremos ser o máximo possível e no eixo $x$ o FPR ($f_{FP}$ que é  $1 -$ TNR) que deveria ser pequeno.\n",
        "\n",
        "Porém quando nosso sinal é raro ou estamos mais preocupados dos falsos positivos que dos falsos negativos é preferível usar a curva chamada *PR* (*precision-recall*, ou seja pureza vs eficiência) \n",
        "\n",
        "\n",
        "Em medicina pode ser mais importante conhecer e optimizar em função da especificidade (excepto em epidemias ou doenças muito raras), mas em física de altas energias por exemplo ela é irrelevante. \n",
        "\n",
        "Específicamente para medicina também é importante conhecer a ***prevalência*** de um sinal  (ou doença), quer dizer quão frequênte ela é na população.\n",
        "\n",
        "$$ \\pi_s = \\dfrac{S}{T}$$"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_JY8HrUz0gsG",
        "colab_type": "text"
      },
      "source": [
        "### Exemplo: Classificação binaria de um dígito \n",
        "\n",
        "Quer dizer, queremos que nosso classificador decida se é ou não um número 5, por exemplo."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8TuUGj0jdxQA",
        "colab_type": "code",
        "outputId": "f1095c72-dfa8-4563-ee64-a4949a1587f5",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "# Python ≥3.5 is required\n",
        "import sys\n",
        "assert sys.version_info >= (3, 5)\n",
        "\n",
        "# Scikit-Learn ≥0.20 is required\n",
        "import sklearn\n",
        "assert sklearn.__version__ >= \"0.20\"\n",
        "\n",
        "# Common imports\n",
        "import numpy as np\n",
        "import os\n",
        "\n",
        "# to make this notebook's output stable across runs\n",
        "np.random.seed(42)\n",
        "\n",
        "# To plot pretty figures\n",
        "%matplotlib inline\n",
        "import matplotlib as mpl\n",
        "import matplotlib.pyplot as plt\n",
        "mpl.rc('axes', labelsize=14)\n",
        "mpl.rc('xtick', labelsize=12)\n",
        "mpl.rc('ytick', labelsize=12)\n",
        "\n",
        "from sklearn.datasets import fetch_openml\n",
        "mnist = fetch_openml('mnist_784', version=1)\n",
        "mnist.keys()"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "dict_keys(['data', 'target', 'feature_names', 'DESCR', 'details', 'categories', 'url'])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 1
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mfJa5INK2C88",
        "colab_type": "code",
        "outputId": "cacab6c2-7790-4fab-d7ef-39ef26718611",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "X, y = mnist[\"data\"], mnist[\"target\"]\n",
        "print(X.shape,y.shape)"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(70000, 784) (70000,)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gUdk6QPB2H8i",
        "colab_type": "code",
        "outputId": "cbc69b45-3e87-4ee8-fd53-618dc5eef810",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        }
      },
      "source": [
        "some_digit = X[0]\n",
        "some_digit_image = some_digit.reshape(28, 28)\n",
        "plt.imshow(some_digit_image, cmap=mpl.cm.binary)\n",
        "plt.axis(\"off\")\n",
        "plt.show()\n",
        "print(y[0])"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAOcAAADnCAYAAADl9EEgAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nocQJocQJocQJocQJof4DO14Dhyk10VwAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "5\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "L_4tIo1Z2VNo",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#algumas definições úteis\n",
        "def plot_digit(data):\n",
        "    image = data.reshape(28, 28)\n",
        "    plt.imshow(image, cmap = mpl.cm.binary,\n",
        "               interpolation=\"nearest\")\n",
        "    plt.axis(\"off\")\n",
        "\n",
        "\n",
        "def plot_digits(instances, images_per_row=10, **options):\n",
        "    size = 28\n",
        "    images_per_row = min(len(instances), images_per_row)\n",
        "    images = [instance.reshape(size,size) for instance in instances]\n",
        "    n_rows = (len(instances) - 1) // images_per_row + 1\n",
        "    row_images = []\n",
        "    n_empty = n_rows * images_per_row - len(instances)\n",
        "    images.append(np.zeros((size, size * n_empty)))\n",
        "    for row in range(n_rows):\n",
        "        rimages = images[row * images_per_row : (row + 1) * images_per_row]\n",
        "        row_images.append(np.concatenate(rimages, axis=1))\n",
        "    image = np.concatenate(row_images, axis=0)\n",
        "    plt.imshow(image, cmap = mpl.cm.binary, **options)\n",
        "    plt.axis(\"off\")"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uQ-Br5y72nAZ",
        "colab_type": "code",
        "outputId": "2686359d-d98b-417d-d82c-38131a5ad3cf",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        }
      },
      "source": [
        "plot_digit(X[3])\n",
        "plt.show()\n",
        "plot_digits(X[0:10])"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAOcAAADnCAYAAADl9EEgAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nE6LECVF/Aepbi9y1gU/6AAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAV0AAAAvCAYAAABAFRnJAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0\ndHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAATB0lEQVR4nO2daVAUVxeG324BBwGBFCIa5UMDSolG\nMcYl4jIaosaFRUUtIYIRFA1CJW4lagJENKiJhRUjEgoVlUWjImphcEXjEgIoiOVaKC6gQFhcgBE5\n3w9qOkxYZKC7seJ9qroKprvve6Z75vS59557hiMiMBgMBkMe+LY2gMFgMN4lmNNlMBgMGWFOl8Fg\nMGSEOV0Gg8GQEeZ0GQwGQ0Z03rCfpTYwGAyG9nCN7WCRLoPBYMgIc7oMBoMhI8zpMhgMhowwp8tg\nMBgy8s443fT0dKSnp8PLyws8z8PLywsZGRltbRbjHcXf3x/+/v7gOA79+vXD/fv329okhpaMGTMG\nY8aM0f5EImpqaxHV1dVUXV1NxcXFwhYUFETLly+n5cuXk7OzMz169IhmzZpFAEihUJBCoaDvvvuu\npZJNkpmZSSYmJmRiYkI8zwubqampJHot4cSJE9S5c2e6ceMG3bhxQ1btkJAQ4jiOUJutQmfOnJFV\nvy0pLy+n8vJyevz4MUVGRtLatWupsrJSUs3c3FwyNTUlU1NT4jiOeJ6n5ORkSTUb4ubNm5SdnU1b\nt24lAMRxXIObi4sLubi4UFVVlSR2qFQqUqlUdObMGRo2bJgkGmITEBBACoWCfHx8GjukUb/6ppSx\nZpGXlweVSoULFy7g/PnzKC0tBQDs37+/weO7d+8OPz8/HDx4EEZGRujfvz8AYNSoUWKYo8Gff/6J\nqVOnoqysDADAcRw6duwIPT09FBUV4eLFiwCAjz76CHp6elq1nZqaiuLiYri4uLTazrS0NAwaNKjV\n7WjLjh07sH79erRr1w6vX78GUHuN/uvk5uYiLCxMuP/Z2dnCvoKCAoSHh0um3alTJ+GznpiYKJlO\nQ1y7dg07d+4EAOzbtw81NTV49OgROI5r9L6rbVywYAE2b96Mjh07imqT+rs5evRoWFhYoKCgABYW\nFqJqiMWKFSsAANu2bYOuri7Gjh2rdRutcrqZmZkAasNs9YVrinbt2gEAvv/+exgYGGD27Nno2rUr\nTE1NAQC9e/dujTkavHz5EhkZGXB3d8fjx4819tnY2GDZsmWYMWMGhg8fLti0cuVKrTTOnDmD27dv\nt9rp1tTUIDc3F3l5eSCZq77dv38fVVVVsmhdvnwZABATE4PU1FRcu3YNALBp0yYAQNeuXXHu3Dl4\neHhgyJAhkthw48YNbN68Gbt370ZFRYVwvS0tLWFkZITr168jISEBCxcuBADY2tqKboOBgQH+97//\nid5uc1i5ciWOHj3aonN37tyJuXPnwsHBQWSr/qGgoOCtdrqXLl0CAKhUKjg4OMDNzU3rNt6ZMV0G\ng8F4G2hVpKt+WpuZmTUa6Q4ZMgSmpqY4ffq00H338PBojWyzmD9/Pvbu3dvgvvT0dDx//hyjRo3C\nmTNnAGh2L5vLzp078cknn7TGTABAfn4+tm/fDg8PD0kiq8Y4ceKE0I22tbXFkSNHAACdO3cWXSs+\nPh7+/v4AgMLCQhARRo8ejaKiIixZskQ4johQVFSEuLg40bTLysqwfPlywY7y8nJhX69evQAAx48f\nh0qlgq2tLQoLC1FUVCSa/r8pLS3F1atXJWu/KRwdHTUiXXNzc3z55ZeoqakBz/8Tg124cAFnz55t\nCxPbjNTUVKxduxaxsbF477336u2PjY0V/IS1tTU2btzYIp1WOV21YRs2bEBSUhLs7e2xePFiYf+A\nAQNw4sQJGBgY4Nq1a5KOk6lJT08HABw5ckToOo4ePRqTJk0CACxZsgRdu3aFvb298DAA0KJufU1N\njSg2z5s3D0DtsIccnD9/HgDg6ekpOKClS5dK0uWtrq5GWloavL298eLFCwC1Y/erV6+Gg4MDqqqq\nhC7a8ePHAUD0se2DBw8iMjKy3uvW1tZISUkBUDvPcPv2bVF1G+Ply5f1shXS0tJga2sr+bCDr68v\nnJ2dhf91dXUb7MqXl5ejb9++ePTokfCas7MzPv74Y0ntA4CKigrJNRrCx8cHt27dwvXr1xscQlm7\ndi3+/vtvAMCvv/4qzEVpTVOzbNrM5pWVlVFNTQ15e3sLs5579uzRck6wdTSUpTBp0iR69uwZJSUl\nUVJSEoWGhtLTp0+Fc9S2GhoaUnp6erO1rl69Svr6+uTu7t5qu4cOHUoA6OLFi61uqznMmzeP5s2b\nJ7x3pVIpmVZ0dLRwL8aPH0/jx4+nsrIyYX9MTIxGRomlpaXG/RGDCRMmaGj07NmTZs6cSffv39c4\n7vDhw8Ix586do3PnzolqR12Cg4MpODhYyF7geZ62bNkimZ62JCQkkKGhocZ18/f3l0SrsLCQCgsL\nheyZtroOAwYMIJ7nKSUlpd6+zMxMMjIyEr4ze/fufVNzjfpV0VPGlixZovFlfv36dUua0ZqbN2/S\nrFmzBG1zc3Pq378/7du3r8nz1MfzPE+zZs1qtt66desIQKudbkFBAVlYWBAAysvLa1VbzaGwsFB4\nzzo6OmRmZkYnT56URCswMFC4tn5+flRWVqbhcImIbG1tNb7Yhw4dEt2OR48e0bfffkvffvst/fHH\nH/TkyZMGj4uMjJTN6ap5G51ubGwsKZVKjfvC83y9eycWJSUlVFJSQiYmJgSAAgICJNFpilWrVpGO\njg7Z2dnVe+g/f/6cZs6cSTzP07Bhw2jYsGGkUqne1GSjfpVNpDEYDIacNOWRW/LEeP78OSmVSlIq\nlcRxHB0/frwlzWhFZWUlTZ48mXieJ2NjYzI2Nqbk5GQqKiqiBw8eNHlu3UjXwcGh2Zqenp4EgNat\nW9cq293d3QkA9e7dm0pKSlrV1pvIzc0le3t7jUg3KChIdJ2goCAKCgoijuNIoVCQk5MTvXz5UuOY\niooKSkxMJH19fcGeNWvWiG6LNnh5ecke6aLOooS2jHRjYmLIzs6O7OzsSKFQaES4AwcOpIEDB9a7\nh2IzefJk2SPdvLw8ysvLo86dO5NCoWhwYZCPjw/xPE/du3fXpmlpF0fUxcDAQJi0GDhwILy9vaFU\nKjFo0CAsWrQIgPjJ9xkZGcKMrDqRW4qFFg2h7cRCeXk5kpOTsXv3bgDA77//DgBYtWoVTExMRLev\nLsnJyRpZGmPHjhUyCsSitLQUW7duBVB7n8eNG4dDhw5pHHPnzh3Mnj0bf/31FwBg+vTpAIBly5aJ\naktThIeH48WLF8IEKsdxQt7w8OHDMWzYMFnsaGpRghTcu3cPMTExAGqzV9ScO3eunh0dO3bEDz/8\ngM8//xwAoK+vL5udcpCdnQ1XV1cAtRk1ixcvruc3Nm7ciB07dgAAAgMDRdEV3ekCwAcffACgdrWT\nl5cXdu3ahV27dgmz11988QW6dOkimt7XX38tpCBp62zVX7p//91c1LOZdbl69Spqampw8uRJPHz4\nEEBtMvWePXtQU1MDfX19Ifm/ffv2ePXqleSr0Q4dOiSsphkxYgSA2pQ3Y2NjUXVUKhUKCwuF/8PD\nw/H06VNER0cLD8ScnBw8e/YMHMeB53m4u7sDqH1gS8XLly8F7eDgYOEhXdfpArULNKKjo4WFPP8l\nsrOzMWXKFOTl5TXr+JEjR8LHx0diq+pTXFwsafvV1dXYvXs35s6dq3H/L168iNDQUHzzzTfC93rf\nvn0gIsyZMwfz588Xx4CmwuAWxev/IisrixwdHTXWcvv6+tLDhw/FaJ6SkpKE7tBPP/2k9fl1hxcW\nLVrU7PN8fX2J4zgyNTWlAQMGaGzqOga6urrCcMfQoUMpICCAYmJi6MGDB8J6c3Nzc9LV1dXabm3I\nzc3VuP6enp7k6ekpiVZJSQlZWFiQhYUF8TyvMVGk3rp3707dunUjnufJwsJCEjvUqFQqunz5MnXr\n1k3QNDQ0pPfff5+mT59ORkZGZGRkJNhmYWFBYWFhVFVVJVmtgbrIOZGWlZVFVlZWDdZXQCO1F44e\nPSqpTXVRDy8YGxtLqlM3Y0b9Pnv16iW8NmTIEI3PSws/o/JlLzRESUkJ7dq1S+ONfvrpp6K0nZCQ\nIFyYx48fN/u8yspKWrFihXDRHR0d6dmzZ1ppr1+/nqZMmdLgFhUV1WQKWEREBEVERBAA6tmzp1a6\n2rJgwQLS0dERNqmL6ly6dIkuXbpEZmZmwgd66dKllJOTQzk5OZSfn0+jRo0inuclHb+rqqqixMRE\nDYcfHBwsjNUWFxdT//79qX///vUeDHFxcRQXFyd58Zu6zm769OmSahER3bt3j0JCQigkJITS0tIo\nOzu73hYQECBcBzmd7o8//ii5042LiyMdHR1SKBRkYWFBp06dolOnTlFmZqZGxkbdYExHR4e6detG\nd+7c0UaKZS8wGAzGW0FTHlnEBwwREenp6ZGenh5xHEd6enp0+vTpVrepjnStrKyafU5lZSWtWrVK\nSMa3tLSUvbSem5sbubm5EQBatmyZJBqZmZmUmZlJPXr0EKJcV1dXSbS04ezZs0KEFx4eLomGSqWi\nFStWCJHLxIkTaeLEiUKGyNOnT2nQoEFCRKNQKGjNmjXk4uKiEfGOGzeOTp48SRkZGZSRkSG6nf8e\nflH3BtqS0tLSNol09+/fTwBIX1+f7t27J4mGUqmkHj16UFRUVL19OTk5NHz48HqRLs/z5OHhoa2U\nfNkL/yYrKwv79+9HWloaXr16Jbzep08fjBw5UjSdKVOmNOu4K1euICwsDPHx8XBycsKBAwdEs6Gl\n1F2WKSafffYZAKCkpARAbR0MdVm/tqSiokKYtZ85c6aobavLU65evRobNmyAoaEh1q1bh1mzZgEA\nTExMkJaWBj8/P2RkZAi1F3755RcolUqUl5fjwoUL2LNnDwDg8OHDcHR0BFBbiSw3N1dUexcsWICI\niAjh/+3btwMANm/eLKqONqiXY8uNjk6tOyIiySrfOTk5wdXVFd27d6+3r6ioCDk5OQAg1P7o27cv\nAKBbt27iGdGUR27hw0QYM1y0aBF16dKl3uC8rq4uTZgwoaXNaxAfH08cx5GlpWWTx23atIk2bdpE\nJiYmxHFcS55colI30r1w4YIkGnXzcXV0dJqzdFE21FGE2Et+t27dSlu3biWe58nIyIhiY2OpuLiY\njh07RseOHaPp06eTgYEB8TxPQUFBQp5mY+zdu1eIkm/duiWqrURE4eHh9Zbair3cVqVSUVJS0hvz\nbKOioigqKkpjYlHOSJeodoUiAPL19ZVVt7S0lBYtWkQcx5GNjY0YTcozkZafn0+bNm2iHj16UI8e\nPeo528GDB9PgwYMpMTFRjDdFRP8ML+jp6ZGfn5/Qpc7Ly6OEhASaPHkyWVpaCjZYWVnRzJkzZatz\n0Bh1ne7OnTtFb1+9eAN1Jmqk6rJpS3JysmROt27mhL6+Ptnb21Pv3r3rTZSFhIRQdXW1qNotxcbG\nRiOLAIC2kzYNkpqaSqmpqTR+/Hjieb7Rh0txcTHFxMTUq1tiaGhIp06darUd2uDv708dO3akiooK\nWXVDQ0OFCfk3LahqJtIOLzx58gQ5OTn46quvcOPGjXr7hwwZgmXLlsHJyQkANErIiUV1dTV+/vln\n4dcqjI2NcevWLWG/ugTjmDFjEBwcLLp+axCrWpmaK1euICUlRcg9bd++PRYuXChJycaWcPfuXcna\nVlfMevr0KaqqqoQSihMnTgRQm3vq7OwMKyurtyYX187OTpJr4ufnB+CfsqVhYWEwMjKqd1xKSgrS\n09M1FkeMHj0aCxcuhFKpFN2uN8FxnNa/4tJS1NXeIiMjwfM8fHx8xB1KaACWvcBgMBhy0lQY3FTs\nXFxcTNOmTaNp06aRtbV1g4nVDg4OdPDgQUnXbD948ICGDh3aYH4dz/Nkbm5Oixcvlky/pdQdXmji\nx+1axOnTp0lHR0e4FlLnAWtLVlaWMOwh9vCC+ocmd+3aRQEBARQaGkoFBQWyLXZoCceOHav3+RVj\neKGxHOSGNo7jhKEZHx8f2bv3avz9/QkA/fbbb7Lo2djYkI2NDfE8T3PmzBGzafGGFy5fvoywsDCk\npaUJS1zr0qFDB6GQeWBgoKRLO4HaWcUDBw4gIiICISEhGvv8/f3h6+srW3FwRvPo168fbGxscPfu\nXdy9exedOnUSrW1199nDw0OWXygRgz59+qBPnz64fv26qO1GR0cDALZs2dJo1oq1tTU6dOiAESNG\nwNvbG0Dt/Wkr4uPjoVAo0KdPH1n0PD09AdRmuzQ3A6rVNOWRG3Lfy5cvrxfR2tnZ0YoVKygwMFDy\nSln/FaKjoyk6OlqSSDc/P59Gjhz51ka6RP8UN1cqlW9Fbup/mcrKStq2bRuZmZkRz/Pk6upKrq6u\ntG3bNsrPz29r8zSYMWMGffjhh2/NpG8raNSvctR0kRd5f5qW8c5QXl4ONzc3pKSkYOrUqQBqIzOp\ne0YMhkw0WjqOTaQxGAyGjLBIl9FmlJeXIzAwUKi/m52dLdtYHoMhMY1GuszpMhgMhvg06nTflL0g\nX0l7BoPBeAdgY7oMBoMhI8zpMhgMhowwp8tgMBgywpwug8FgyAhzugwGgyEjzOkyGAyGjPwfxTQp\nkj+71OsAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "38JLTMRB22tZ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.model_selection import train_test_split\n",
        "X_train, X_test, y_train, y_test = train_test_split(X,y,test_size=10000,random_state=42)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "yyRclR1-5ctd",
        "colab_type": "text"
      },
      "source": [
        "Agora redefinimos o alvo ($y$) para ter um resultado binário: Sim ou Não (ou *verdadeiro ou falso*). O sinal (ou Sim) é o dígito 5."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "23VoNUID4biN",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "y_train_5 = (y_train == '5')\n",
        "y_test_5  = (y_test == '5')\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7xBNKii_7Gyg",
        "colab_type": "text"
      },
      "source": [
        "Chamamos agora um modelo de ***Stochastic Gradient Descent*** para classificar os dados. Por default é um SVM,  [vide documentação](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.SGDClassifier.html#sklearn.linear_model.SGDClassifier)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bW5fgZ1b47wS",
        "colab_type": "code",
        "outputId": "9ee28a22-a36b-4b9b-ccc5-c60ea5259eff",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 119
        }
      },
      "source": [
        "from sklearn.linear_model import SGDClassifier\n",
        "\n",
        "sgd_clf = SGDClassifier(max_iter=1000, random_state=42)\n",
        "sgd_clf.fit(X_train, y_train_5)"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "SGDClassifier(alpha=0.0001, average=False, class_weight=None,\n",
              "              early_stopping=False, epsilon=0.1, eta0=0.0, fit_intercept=True,\n",
              "              l1_ratio=0.15, learning_rate='optimal', loss='hinge',\n",
              "              max_iter=1000, n_iter_no_change=5, n_jobs=None, penalty='l2',\n",
              "              power_t=0.5, random_state=42, shuffle=True, tol=0.001,\n",
              "              validation_fraction=0.1, verbose=0, warm_start=False)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9In76SR48mvP",
        "colab_type": "code",
        "outputId": "3a37ec0f-f315-4b8a-f08e-1dc257764d5f",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 513
        }
      },
      "source": [
        "print(sgd_clf.predict([X_test[15]]))\n",
        "plot_digit(X_test[15])\n",
        "plt.show()\n",
        "print(sgd_clf.predict([X_test[1]]))\n",
        "plot_digit(X_test[1])\n",
        "plt.show()"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[ True]\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAOcAAADnCAYAAADl9EEgAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nJr/Aa0b4AAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "[False]\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAOcAAADnCAYAAADl9EEgAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n3wAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vIZJYr8g9VyY",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "y_predict_5 = sgd_clf.predict(X_test)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "oCdzhykdtPC8",
        "colab_type": "text"
      },
      "source": [
        "Agora podemos testar a acurácia do modelo com cross-validation, nesse caso com 3 sub-conjuntos aleatórios dos dados de treinamento, e também a validação do conjunto de teste."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "G-Eit0CV9g9u",
        "colab_type": "code",
        "outputId": "b2808651-340b-463d-ed48-0d51b23ce842",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "from sklearn.model_selection import cross_val_score\n",
        "from sklearn.metrics import accuracy_score,precision_score,recall_score\n",
        "\n",
        "print(cross_val_score(sgd_clf,X_train,y_train_5,cv=3,scoring=\"accuracy\"))\n"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[0.9629  0.96775 0.96435]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "s78pzz63zn0h",
        "colab_type": "code",
        "outputId": "2bcc44c0-e77e-4a9b-cb72-092bcc4f5121",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "print(\"acuracia = {:2.3f} \".format(accuracy_score(y_test_5,y_predict_5)))\n",
        "print(\"precisao = {:2.3f} e eficiencia = {:2.3f}\".format(precision_score(y_test_5,y_predict_5),recall_score(y_test_5,y_predict_5)))"
      ],
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "acuracia = 0.966 \n",
            "precisao = 0.853 e eficiencia = 0.767\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VYXkEGfrtNbl",
        "colab_type": "text"
      },
      "source": [
        "Para comparação, podemos criar um classificador que nunca responde sim. A classe não faz nada no fit  e na predição vai dar sempre a resposta negativa. Qual será a acurácia desse classificador?"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "InVptWl_tLnq",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.base import BaseEstimator\n",
        "class Never5Classifier(BaseEstimator):\n",
        "    def fit(self, X, y=None):\n",
        "        pass\n",
        "    def predict(self, X):\n",
        "        return np.zeros((len(X), 1), dtype=bool)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6VDtrT_PIFBh",
        "colab_type": "text"
      },
      "source": [
        "Chamaos a validação do modelo e tanto na cross-validation como na validação no teste da uma acurácia de ~90%"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9OM3QZzot5zK",
        "colab_type": "code",
        "outputId": "c7640c35-b72d-46f8-f3b3-8263c75f38b0",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "never_5_clf = Never5Classifier()\n",
        "print(cross_val_score(never_5_clf, X_train, y_train_5, cv=3, scoring=\"accuracy\"))\n",
        "y_pred_never_5=never_5_clf.predict(X_test)\n",
        "print(accuracy_score(y_test_5,y_pred_never_5))"
      ],
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[0.9119  0.90875 0.91055]\n",
            "0.9063\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AOb55uncIfFS",
        "colab_type": "text"
      },
      "source": [
        "Mas se pesquisamos a precisão e eficiência, elas são zero (ou seja todas as instâncias negativas são classificadas corretamente, mas nenhuma das instâncias positivas)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dP2-UwMiycX5",
        "colab_type": "code",
        "outputId": "8f3cce7b-bf07-4bc3-d265-b469ea21d696",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 88
        }
      },
      "source": [
        "print(precision_score(y_test_5,y_pred_never_5),recall_score(y_test_5,y_pred_never_5))\n"
      ],
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "0.0 0.0\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/sklearn/metrics/classification.py:1437: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
            "  'precision', 'predicted', average, warn_for)\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "36r8vh5-u3Rm",
        "colab_type": "text"
      },
      "source": [
        "Recapitulando, a predição de \"chutar\" todas as predições como não sendo o dígito 5 dá certo ~90% do tempo pois ~1/10 dos dígitos nos dados são 5, então reflete a *prevalência* do sinal. Mas se examinarmos a precisão (pureza) e a eficiência (*recall*) vemos que esse mêtodo não tem boa performance.\n",
        "\n",
        "\n",
        "## Matriz confusão com o SciKit Learn\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KjqGsv19uAGz",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.model_selection import cross_val_predict\n",
        "\n",
        "y_train_pred = cross_val_predict(sgd_clf, X_train, y_train_5, cv=3)\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mCXwq6OHwsXI",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "y_train_predallzero = never_5_clf.predict(X_train)\n",
        "y_train_predperfect = y_train_5 "
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kglv-yf-wowN",
        "colab_type": "code",
        "outputId": "944718e3-a46e-4b5d-d7ae-d09e0ea4c3e2",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 153
        }
      },
      "source": [
        "from sklearn.metrics import confusion_matrix\n",
        "print(\"Test sample confusion matrix\")\n",
        "print(confusion_matrix(y_test_5,y_predict_5))\n",
        "print(\"\\n\")\n",
        "print(\"Training sample confusion matrix\")\n",
        "print(confusion_matrix(y_train_5, y_train_pred))"
      ],
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Test sample confusion matrix\n",
            "[[8939  124]\n",
            " [ 218  719]]\n",
            "\n",
            "\n",
            "Training sample confusion matrix\n",
            "[[54103   521]\n",
            " [ 1579  3797]]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "j3dCqXwtyM6e",
        "colab_type": "code",
        "outputId": "960df5ff-b31c-44cd-c2fe-f4ae5d92baee",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68
        }
      },
      "source": [
        "print(\"Perfect score confusion matrix\")\n",
        "print(confusion_matrix(y_train_5, y_train_predperfect))"
      ],
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Perfect score confusion matrix\n",
            "[[54624     0]\n",
            " [    0  5376]]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hIky12W9yQ2H",
        "colab_type": "code",
        "outputId": "6794f825-cc6f-4a65-f010-f66b25fd69c3",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68
        }
      },
      "source": [
        "print(\"Never 5 confusion matrix\")\n",
        "print(confusion_matrix(y_train_5, y_train_predallzero))"
      ],
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Never 5 confusion matrix\n",
            "[[54624     0]\n",
            " [ 5376     0]]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1nqfC124qBN8",
        "colab_type": "text"
      },
      "source": [
        "O resultado do valor da precisão e da eficiência para o modelo SGD é"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZQSB0W-byW6d",
        "colab_type": "code",
        "outputId": "68530a55-10fe-400d-f5e0-4a8bc0346902",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "print(\"precision {:2.1f}%\".format(100*precision_score(y_train_5, y_train_pred)))\n",
        "print(\"recall {:2.1f}%\".format(100*recall_score(y_train_5,y_train_pred)))"
      ],
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "precision 87.9%\n",
            "recall 70.6%\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "YAQYm-xqqg9x",
        "colab_type": "text"
      },
      "source": [
        "Para acessar os valores da função de decisão:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "etvyhBAg1NXn",
        "colab_type": "code",
        "outputId": "e73c2a83-ce18-4e69-aed9-0d74b3fe21e2",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "y_scores = sgd_clf.decision_function(X_train)\n",
        "y_scores[0]"
      ],
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "-14144.506062058408"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 22
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Jv-6ADr5q0sP",
        "colab_type": "text"
      },
      "source": [
        "A função `precision_recall_curve` retorna os valores de pureza e eficiência do método a medida que muda o *threshold* da decisão"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jF28YqGx1qu3",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.metrics import precision_recall_curve\n",
        "\n",
        "precisions, recalls, thresholds = precision_recall_curve(y_train_5, y_scores)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PgCBItH0rORR",
        "colab_type": "text"
      },
      "source": [
        "Podemos salvar o ponto limite onde a precisão/pureza alcança 90%. Neste ponto, quanto é o valor da eficiência?"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KoxjuUlQ3K4L",
        "colab_type": "code",
        "outputId": "b161d0e5-0a6a-4440-c104-2ff8d4207515",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "threshold_90_precision = thresholds[np.argmax(precisions >= 0.90)]\n",
        "print(threshold_90_precision)\n",
        "\n",
        "recall_90_precision = recalls[np.argmax(precisions>=0.9)]\n",
        "print(recall_90_precision)"
      ],
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "483.1147511086346\n",
            "0.7345610119047619\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "yHWhiBAhrhEB",
        "colab_type": "text"
      },
      "source": [
        "A função definida embaixo ajuda a fazer o gráfico da evolução da pureza e recall com o valor do *threshold* de desição. O *threshold* de 90% de pureza está indicado em vermelho."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "I7lLAIvy1rgk",
        "colab_type": "code",
        "outputId": "62d24459-4ca5-45cb-ee6f-c673fde9008b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 292
        }
      },
      "source": [
        "def plot_precision_recall_vs_threshold(precisions, recalls, thresholds):\n",
        "    plt.plot(thresholds, precisions[:-1], \"b--\", label=\"p\", linewidth=2)\n",
        "    plt.plot(thresholds, recalls[:-1], \"g-\", label=\"$\\epsilon$\", linewidth=2)\n",
        "    plt.legend(loc=\"center right\", fontsize=14) # Not shown in the book\n",
        "    plt.xlabel(\"Threshold\", fontsize=16)        # Not shown\n",
        "    plt.grid(True)                              # Not shown\n",
        "    plt.axis([-50000, 20000, 0, 1])             # Not shown\n",
        "\n",
        "plt.figure(figsize=(8, 4))                      # Not shown\n",
        "plot_precision_recall_vs_threshold(precisions, recalls, thresholds)\n",
        "plt.plot([threshold_90_precision, threshold_90_precision], [0., 0.9], \"r:\")         # Not shown\n",
        "plt.plot([-50000, threshold_90_precision], [0.9, 0.9], \"r:\")      # Not shown\n",
        "plt.plot([-50000, threshold_90_precision], [recall_90_precision, recall_90_precision], \"r:\")# Not shown\n",
        "plt.plot([threshold_90_precision], [0.9], \"ro\")                   # Not shown\n",
        "plt.plot([threshold_90_precision], [recall_90_precision], \"ro\")               \n",
        "plt.show()"
      ],
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfkAAAETCAYAAAA1aX8AAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0\ndHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAAgAElEQVR4nO3dd3xUVfr48c+TQkJIQu9FqkFQAUER\nUAmCKDZcUZcVrLgo6m9ld2XXsnbXspZdsVBcXVaxgAhfBAuCGqSIIooCSugQ+tISAuk5vz/ODJmE\nlElyZ24y87x5ndfMvXPuvc+ZmfDMbeeIMQallFJKhZ4ItwNQSimlVGBokldKKaVClCZ5pZRSKkRp\nkldKKaVClCZ5pZRSKkRpkldKKaVClCZ5pZRSKkT5leRF5G4R+V5EckRkWgV1/ygie0UkQ0TeFJEY\nRyJVSimlVKX4uye/G3gSeLO8SiJyMXAfMBg4BegIPFadAJVSSilVNX4leWPMbGPM/wEHK6h6E/CG\nMWadMeYw8ARwc/VCVEoppVRVRDm8vu7AXJ/pn4DmItLYGFPsB4KIjAXGAsTGxvZu166dY0FsO76N\n3IJcx9anVG0hIghin3sfxT5GEIGIEEEEERJhXxdbz/taseU8jxESUWy9EUQQKZEnzReE6IjoE9P+\nKCwsJCIifC8NCqX25+VFEB1dWObrOTkRFBYKsbEFeL5qrrR/795YMjKiy3y9Q4djJ9qxe3ddMjNP\nTpOnnnrUkVg2bNhwwBjT1JGVlcHpJB8PpPtMe58nUOIogDFmKjAVICkpyaSmpjoWxMcbPuZg1kHy\nC/PJK8grtY6h7D77y+rPv6xlyuv/359lNm7aSJfOXaq8nZrUlqrEtWXLFjp27Oh4XHkFeeQU5GCM\nOVGnKs+92/D3eV5hHkeyj1BgCig0hSeKMabYdKEppMAUsPfQXuLqxdnpwoIT8711cgtyycrLOmnZ\nksX4/HNT64TWNI5rTJ3IOkRHRBMdGU39mPo0jWtKg9gGxEbFkhiTSHydeNK2pNG3R1/qRtclvk48\n9WPqkxCTQOuE1kRHlv0fcahISUkhOTk56NvNyYHcXEhIqNrya9bYdXzzDWzcCC+/bOcfPw5160Jh\nIQwYAElJcNZZ0KQJjBpl6xw7BnFx9vmECevp1asrfftCx46cSP6B9tNPtg2lGT686H1ZsgS2by/+\nukhRW6pLRLZXXKua26jMADUi8iTQxhhzcxmv/wT83Rgz0zPdGDgANCm5J+/L6SRf27j1h15TaPud\nab8xhszcTApMAQWFBcUe8wvzOZpzlOz8bDJzM8nMzSSvMI+8gjzyCvM4mnOU/ML8E8t4n+fk53A8\n77j9weypfzT3KIeyDnE4+zD5hfknSnp2Onsz91JgCqrdlqiIKDo36kzTuKa0iG9B64TWtElsQ5O4\nJsRGxdKsXjM6N+pM2/ptq70tNwX7uz9rFlx7bdF048Zw8KBNyLfcAmPG2Pk//gjffw/z59tkvmAB\nTJgAd94J7dvDeefBsmUnr//xx+GhhyAtDco6OOubcpo2zeHAAXttdpMm0LcvDBliY2zd2pk212Qi\nssoY0yeQ23B6T34d0AOY6ZnuAewrL8ErpZwhIiTEVHHXzCG5BbnsytjFkewjJ34U5BbkcijrEP87\n/j8yczPJysviSPYRsvKz2LhjI/Ua1iMr3847lnuMQ1mH2JO5h/UH1rOe9eVuLzEmkb6t+9KjeQ/a\n1W/H2a3Ppm/rvidOPYSrtDR4/XX45z9tsl282CbRFi1s2bvX1jvo+Z952TLo3bto+TFjbKL39dxz\ncPrpNsk3awZNm8L//gdt2tj1vPACXH+9rdusmV3n6tWwahWsXQupqfaHgld+Pgwfvov9+zvy7bew\nfz98/LEtf/wjPPEEPPhg8PbuQ5VfSV5Eojx1I4FIEYkF8o0x+SWqvgVME5F3sFfk/w2Y5ly4Sqma\nrE5kHTo07OB3/bL2ZI9kH2H7ke0cyjrE3sy9bE/fzrYj28jKz+JY7jH2ZO5h5a6VZORksHDLQhZu\nWXhi2VYJrWgR34IODTpwSv1TuK77dfRt09eJ5tVoGRk2gWdlFZ//yy82IW/YYPfY774bhg6FPn1g\n50576HrJErjppqJlunQpSvIjR8Lu3XD//XYvG2D2bPtoTOlJOCYG+ve3pSxRUTB69A6SkztiDGzb\nBsuXw4wZ8NVXdm/eu+7Vq6F7d4h24AzO9dfb0wzTp9v3I9T5uyf/N+ARn+nRwGMi8ibwC9DNGLPD\nGPOZiPwD+AqoC3xYYjmllKpQg9gGNGjRoNw6Ofk5bD2ylZ/2/sTa/WvZemQrH2/8mN1Hd7P76G5+\n2PMDAC+ueJGuTbrSu2VvhnYaytWnXU18nfhgNCPgcnJg61Z7aDs/3+5pr1xZvM4tt8ADD0Dnznb6\nwQeLXmvb1pbLLy++zIwZ8P77Fe9FO7WXLQIdOtgyapRtV4ynh5W8PLj6anvOfuJE6Natetvas8f+\noMgNk2uz/UryxphHgUfLeLnYX4sx5kXgxWpFpZRSFYiJiqFrk650bdKV3/JbwJ4u2Hp4K0eyj7D5\n8GZStqXw1k9v2UP/B9bzzpp3uO2j27i99+3c0ecOujfr7nIrKs8Ye/h91Ci7hz16tD0P3qQJ3H67\nvXDsd7+zSbE63DxMHuPThdratbadW7favfleveCdd+C006q27gLPJSMhclNDhcKkmUqpcFAnsg5J\nTZLo26Yv159xPVOvmMrBvxxk+a3LeWbwM5zV8izyCvN4ZeUrnD7pdEbOGsmO9B1uh+23rCzo2hUG\nDbKJD+xh53ffhUaN7Ln0Bx+sfoKvSXr1ghUrIDnZJuYff7QX/i1cWJSwK6PQc5efJnmllAoB9erU\no1/bfvz1vL+yauwqVoxZwagz7D1QM9bNoNPETtwx/w7W7Cvjnqoa4uBBe9Hbhg1F8669Fo4ehUcf\ndSuq4OjZ056nP3IErrgCDh2y1xW8/Xbl1+VN8pGRzsZYU2mSV0qFlb5t+jL96umsHbeWizpeRH5h\nPlNWTeHMyWdyw5wb2Hp4q9shlio+3t5zDvZWNWNg5kw7P1wkJNiL/u6/356/P++8yq9DD9crpVQY\n6N6sO5/f8DlLb1nKyNNHIgjTf55Ox4kdGf7+cLLzs90OEbAXoWVl2fPUM2bAli32HHy4ioqCp56y\n74P3YsJt2+DKK+HAgYqX1z15pZQKIwPaDeC9Ee+x4rYVnNvmXAA+Sv2I0149jY0HN7oa265dEBsL\n771nk31iot2DVcU9/DDMmwf9+tn78stz003w179Cy5bBic1tmuSVUgo4p/U5fDPmG7648QuaxjVl\n25FtnDHpDJ5e8nS53SoHijG2oxmwF9Tt3Bn0EGqNZ5+1F+ht2mTv/7/sMvjgA8gu5WDM3XfDM88U\nvbehTpO8Ukr5uLDDhay/ez3XdLuGnIIcHvjyAW776LagH74/55yi52+8AZ06BXXztUrLlvD113Dv\nvfbIxyefwHXXwRln2Pvsw5kmeaWUKqFR3UZ8cO0HvDzMjrzy5uo36TG5R9AuyuvXz/YdD/Yis1tv\nDcpma7X4eNv17vbt8K9/2XkFBfbuA19Ll9rb744dC36MbtAkr5RSZbj7nLv5fPTntIxvyYaDGxg4\nbSCbDm0K6DZXr7b3hYMdxe3vfw/o5kJOs2Zwzz0wd67t0rdRo+Kvjxljb79LS3MnvmDTJK+UUuW4\nqNNFrLtzHb1a9CItI42zXz+br7d/HbDt+V71vX69DtBSVVdeaQ/dl6RX1yullCqmYd2GfHHjFwxo\nO4Aj2UcY+vZQvt/9vePbyc21t4X99JPti15V38SJdkQ77+F57fFOKaXUSRrWbciiGxcxPGk4OQU5\nXD3jarYc3uLY+rOzbQ92U6faC8bCZU8z0F580d5i94Mdr0g7w1FKKVW62KhY3r/mfXo070FaRhqD\n3xpMVl5WxQuW5513oH17YupG8NJH7fl2/Du4cMdeyPrd7+zjU0/Z2xL1cL1SSqkyxUbFsujGRbRK\naMW2I9v444I/VnldzRYtgrFjYft2BEN7tvN2zFgi3nvHwYjD2/jxtjvczz6zV93rnrwL4tLSYNo0\nO5GXZ4cbmj7dTh8/bqdnzLDT6el2evZsO33ggJ2eN89O791rpz/7zE6npdnpRYvs9JYtdnrxYjud\nmmqnly+302vX2mnvoMyrV9vp1avt9MqVdnrtWju9fLmdTk2104sX2+ktnsN4ixbZae+lnJ99Zqf3\n7rXT8+bRc/z4ov4YZ8+2r6en2+kZM+z08eN2evp0O+29+XPaNDvt9frrMGRI0fRrr8GwYUXTL71k\nr0jxev55GDGiaPqZZ2DkyKLpJ56wY1l6PfywHaDa6/777X9SXvfeC3fdVTQ9frwtXnfdZet4jR1L\nh9dfL5q+5Ra7Da/Ro20MXiNH2hi9RoywbfC68krbRq9hw+x74DVkiH2PvJKTXf/uJXq/Sy5890hO\ndvW71933s3bhu8f99xdNV+K71ySuCWsWncqE5RFMWTWFD9Z9UKXvXpeXXy56fz0ic44XH/RdVUvz\n5vDf/9rnEyYUjd6ne/JKKaXK1KhuIy479VIARn44kp0Zuyq9jqiMjNJf2FF7hr+tDX7zG/jLX+xe\nfHS0vbCxaVO3owoOcaO7xpKSkpJMqndvJAylpKSQ7LtHFGa0/eHb/tre9kJTyOjZo3lv7XtERUSx\n6IZFDGw/0O/ls5q3oO7+fSe/cMopdtSVEBfMzz831/YieOQIfPwxdO8elM2WS0RWGWP6BHIbuiev\nlFJVFCERTL96Opefejn5hfkMf384245s83v5z5Pv4RhxxWfGxWkPOAFQp449I/XzzzUjwQeLJnml\nlKqGCIlg9nWzOb3Z6aTnpHPtB9dSUFjg17KJd/Rj/5NTyW3Y3PZ6c8op9h66UaMCHHV46tjRjuQX\nTjTJK6VUNUVHRrNg9AIaxjbk+93f88TXT1S8EDavd3hwFHXO7AoXXGAP0WuCVw7SJK+UUg5oldCK\nKZdPAeDxxY/z876fy62/bRtF98NPmWKLUg7TJK+UUg65tvu1jOk1BoPhxW9eLLNeejqceSaMG3eW\nHSUtKckWpRymSV4ppRw0of8EIiWS//70XxZsWlBqnY4d7RCodesWkJCA7ePA23+CUg7SJK+UUg5K\napLEIwMfAWD8gvEnXYRXWAiHDtnnF1/suX3ukUdsUcphmuSVUsphEwZMoGV8S9YfWM/c1LnFXvN2\nqAgwdKin98E337RFKYdpkldKKYfFRsUy/lzbpe6UVUUX1BkDN91kn99+u0//6R072qKUwzTJK6VU\nAIzpNYaEOgl8vvlzVu6y4xH4doX/17/6VF60qGiMA6UcpEleKaUCoHFcY8b0GgPAy9+9DMA119jX\nkpOhQwefyk8+aYtSDtMkr5RSAXLXOXZUvHfXvMuujF1062YvvPvqqxIV337bFqUcpkleKaUCpHOj\nzlzW5TIKTMGJXvBESqnYtq0tSjlMk7xSSgWQ93a6qd+9xb1P7CA3t5RKn31mi1IO0ySvlFIB1KdV\nH3rVuxwTmcXUDY8SHV1KpWeesUUph2mSV0qpABIR2v3yTzBCZod3OZh14ORK779vi1IO0ySvlFIB\ntvbrzrBxGCYyh0krJ51coUULW5RymF9JXkQaicgcETkmIttF5Poy6sWIyGQR2Scih0Rknoi0djZk\npZSqPT77DDZvBr75MwCvrHyF7Pzs4pXmzbNFKYf5uyf/KpALNAdGAZNEpHsp9e4B+gFnAq2Aw8DL\nDsSplFK10t/+Zh+bZA6iZ4ue7D+2n1m/zCpe6YUXbFHKYRUmeRGpB4wAHjLGZBpjlgIfATeUUr0D\nsMAYs88Ykw3MAEr7MaCUUmEhMtI+/v1JYVyfcQC8uvLV4pVmzbJFKYeJMab8CiK9gGXGmDifefcC\nA40xV5So2wd4CbgWOAL8G9hvjBlfynrHAmMBmjZt2nvmzJnVbErtlZmZSXx8vNthuEbbH77tD4e2\nb9wYzw8/NGTYsD1ExqVz1fKryDf5vHvOuyQUJIR8+8sTDp9/eQYNGrTKGNMnkNuI8qNOPJBRYl46\nkFBK3Y1AGrALKADWAHeXtlJjzFRgKkBSUpJJTk72L+IQlJKSgrY/2e0wXBPO7Q+Hthc1rxMAVx2+\nilm/zGJj3EYuMBfY9s+ebatcfbULEbonHD5/t/lzTj4TSCwxLxE4WkrdV4EYoDFQD5gNfFqdAJVS\nKpTc0fsOACZ/P5kC4xlrfuJEW5RymD9JfgMQJSJdfOb1ANaVUrcnMM0Yc8gYk4O96O4cEWlS/VCV\nUqp2GToU/vxnyMwsmndhhwvp0qgLezL3sPKQHZ2OuXNtUcphFSZ5Y8wx7B754yJST0QGAMOB0kZT\nWAncKCL1RSQauBPYbYwppfcHpZQKXRs2wMKF8PrrULdu0XwR4eaeNwPw6V7Pgc769W1RymH+3kJ3\nJ1AX2A+8B4wzxqwTkfNFxOc3KvcC2dhz8/8DLgV+42C8SilVKyQl2cf+/YuusPcafeZoIiSCpQeW\nsjNjJ8yYYYtSDvPnwjuMMYeAq0qZvwR7YZ53+iD2PnqllFLA2WefPK9d/XaMOG0EH/zyAf/58T88\nNOkL+8Jvfxvc4FTI025tlVLKYXv3Fj2///7S69zS8xYAJn0/ieyPZsMnnwQhMhVuNMkrpZTDvKPG\nxsXZUpqLOl1Es5hm7Mncw5f7VpRdUalq0CSvlFIOW7DAPj77bNl1oiKiGNZiGADbXnkSpk8PQmQq\n3GiSV0ophz39NHzwAfymgsuOL2lxCZESSfePviF78qvlV1aqCjTJK6WUw9q3h2uugdYVjMHZIrYF\nz/2vF+0PQ8zyFXbBd94JRogqTGiSV0opB6WlweLF/tVttmgR/+/fP3NKOogBtm+HsWM10SvH1Igk\nH5eWBtOm2Ym8PNvZs/f81PHjdtp7D2l6up329vV84ICd9o7FvHevnfZe+ZKWZqcXLbLTW7bYae9f\nYWqqnV6+3E6vXWunV3p6olq92k6vXm2nV66002vX2unly+10aqqdXrzYTm/ZYqcXLbLTaWl2+rPP\n7LT38tt58+g5frxtB9h2JSfbdoJtd3KyfR/Avi/JyfZ9Avu++fb9/PrrMGRI0fRrr8GwYUXTL70E\nV15ZNP388zBiRNH0M8/AyJFF0088AaNHF00//DDcckvR9P332/+UvO69F+66q2h6/HhbvO66y9bx\nGjuWDq+/XjR9yy12G16jR9sYvEaOtDF6jRhh2+B15ZW2jV7Dhtn3wGvIEPseeSUnu/7dS/R+l1z4\n7pGc7Op3r7vvZ+3Cd6/Ype8Offd69IAnn4ScCyv+7nV5+WWisnMp5vhxePBBlHJCjUjySikVCo5n\nweHD9vdVlB+9kERllBz7y2PHDmcDU2GrwqFmgyEpKcmkevdGwlC4j8Sk7Q/f9oda2z/+GC6/HAYM\ngKVLK66f3aIFsfv2nfzCKafAtm2Ox1fThNrnX1kiEvChZnVPXimlHJKSYh8vuMC/+ltuu+3k++Pj\n4uDvf3c0LhW+NMkrpZRDvJdj+F6aUJ79Q4bA1KmYRo0wwLb6cOilZ2GU9g6unKFJXimlHPC//9lr\nImNi7OF6v40ahfTqxU/dm9Dhj/B/vbXnO+UcTfJKKeUA7178wIE20VfKokUse/NRAOZvmO9oXCq8\naZJXSikHnHMOPPUU3H571Za/MsneXrhwy0LyC/MdjEyFM7+GmlVKKVW+pKSyR5yr0Guv0Rbo3Kgz\nmw5t4sc9P3J261LGqFWqknRPXiml3DZvHsybx8BTBgKweLufXeYpVQFN8kopVU0TJ4IIrFhRxRV8\n+il8+umJJP/l1i+dC06FNU3ySilVTXPm2Me5c6u3nsEdBwPw1bavyMzNrGZUSmmSV0qpatu0yT5e\nf30VV/DSS/DSS7RKaEX/tv3Jzs/mk42fOBafCl+a5JVSqhpSU2HnTmjYELp1q+JKvvjCFuCqpKsA\n+GzTZw5FqMKZXl2vlFLVMN9zW/vll0NkZBVX8tFHJ54O6Wi7y/ti6xcYYxCRakaowpnuySulVDV8\n4jmqfumlzqyvR4seNKrbiB3pO9h8eLMzK1VhS5O8UkpV0eHD8KXnQvihQ6uxouefPzE2fYREcGGH\nCwH4dOOn1YxQhTtN8kopVUU5OUXPGzWqxoq++cYWj4s7XQzA1zu+rsZKldJz8kopVWUtWsCxYxAb\nW80VffhhsUnv/fJfb/9az8uratE9eaWUqoa4OIhw+H/Szo060zK+JfuP7Wft/rXOrlyFFU3ySilV\nBbt2QbNm8LUTR9SfecYWDxFhWOdhAMzbMM+BDahwpUleKaWqYMkSO4b8wIEOrGz1alt8XJF0BQAf\nb/zYgQ2ocKXn5JVSqgoWLLCPzz7rwMref/+kWRd2uJCoiChW7FzBkewjNIht4MCGVLjRPXmllKok\nY050UMfFFwdmG4kxifRr049CU8jCzQsDsxEV8jTJK6VUJW3ZAmlp0LgxnHGGAyt84glbSrj81MsB\nPWSvqk6TvFJKVdKyZfbxvPMcurI+NdWWEi7rchlg+7E3xjiwIRVu9Jy8UkpV0tKl9rFfP4dWOH16\nqbO7Ne1Gs3rN2HdsH5sPb6Zzo84ObVCFC79+g4pIIxGZIyLHRGS7iJQ5oKKInCUiX4tIpojsE5F7\nnAtXKaXcl50NderAoEGB3Y6IcG6bcwFYsXNFYDemQpK/B5peBXKB5sAoYJKIdC9ZSUSaAJ8BU4DG\nQGfgc2dCVUqpmuGtt2y/9b17O7TChx+2pRTntrZJ/tud3zq0MRVOKjxcLyL1gBHA6caYTGCpiHwE\n3ADcV6L6n4AFxph3PNM5wK8OxquUUjVCXJyDK0tLK/Olfm3tOYElO5Y4uEF3ZWRksH//furXr8+v\nv4ZmioiOjqZZs2YkJia6GodUdDGHiPQClhlj4nzm3QsMNMZcUaLul8Aa4GzsXvy3wF3GmB2lrHcs\nMBagadOmvWfOnFnNptRemZmZxMfHux2Ga7T94dv+2tj2vXtjado0u+pjx/vwp/05BTlcsewK8k0+\nc/vPJSE6ofobdlFkZCSJiYm0bt2ayMhIoqJC79IwYwy5ubns2rWLjIwMCgoKSq03aNCgVcaYPoGM\nxZ93Nx7IKDEvHSjtm9YGOAu4CJvs/wG8BwwoWdEYMxWYCpCUlGSSk5P9DjrUpKSkoO1PdjsM14Rz\n+2tb242Btm0hMxN+/hnatave+vxt/7k7zmXJjiWYdobkpIrr12SbNm2iVatWxMXFcfToURISaveP\nlvLExcWxe/duOnd274JJf87JZwIljzckAkdLqZsFzDHGrDTGZAOPAf1FpH71wlRKKfdt2WL7rI+O\ntsneMfffb0sZktsnA5CyLcXBjbojLy+PunXruh1GUNStW5e8vDxXY/AnyW8AokSki8+8HsC6Uur+\nDPge/9cbO5VSIcPble3AgeDo6K8HD9pSBu/Qs0t3LHVwo+4Jl6Fza0I7Kzxcb4w5JiKzgcdF5Dag\nJzAc6F9K9f8AH4rIROyPgIeApcaYdAdjVkopV6Sk2MehQx1e8dSp5b58VsuzAFizfw15BXlER0Y7\nHIAKVf7eQncnUBfYjz3HPs4Ys05EzheRTG8lY8yXwAPAx566nYEy76lXSqnawhiYPds+P++84G67\nYd2GdGvajez87JDZm1fB4VeSN8YcMsZcZYypZ4xpZ4x51zN/iTEmvkTdScaY1saYhsaYK4wxZd8b\nopRStcSmTeC9SLprV4dXfu+9tpTDO778wi06WI3yn/Zdr5RSfvCOOgcO9VfvKyvLlnIM6TgE0CSv\nKif0blBUSqkAuP12SE6GwsIArPzVVyuscn6784mKiGLV7lWkpafRtr6Tl/eriiQnJ9O1a1diYmJ4\n6623ALjtttt49tlniXD8V59zam5kSilVg4jYw/Tdurmz/Xp16nFl0pUYDDPXhW/nYW565513KCws\n5JtvvmHKlClMnTqVf/3rX26HVS5N8kopVYGAj/I6frwtFbiu23UAzPwl9JK8SNnF9+aDqVPLr+ur\nd2//6vmrZcuWTJw4ka5du3LdddcxYcIEXnzxxao3Ogg0ySulVAWee84mjFmz3I1jWJdhCMKPe34k\nOz/b3WDC0Lnnnlvs3vd+/fqd6Lq2ptJz8kopVYFFi+CHHyA/P0Ab8POQb2JMIt2admPd/9bx454f\nTwxeEwr8PVoydqwt/li1qurxhArdk1dKqXLk5sKyZfZ5Tehm3zu+/PK05S5HEn6+/fZbfAd1W7Fi\nBa1atXJ9pLnyaJJXSqlyrFwJx4/bC+5atAjQRu66yxY/ePux/2LrF+VXVI7bvXs348ePJzU1lVmz\nZvHcc8/xxz/+0e2wyqWH65VSqhzermwHDgzgRioxYMuFHS4EYPH2xeQW5FInsk6golIljBo1ioKC\nAvr27YuIMGbMGE3ySilVmy1ZYh8Deqj++ef9rtoqoRWnNTmNXw/8yne7vuO8dkHuYzeMRUVF8cor\nr/DKK6+4HYrf9HC9UkqVoaAAlntOfQ8Y4G4svs5vdz6g5+VVxTTJK6VUGQoL4Y034KGHoHXrAG6o\nMpeMAwPa2V8cmuRVRfRwvVJKlSE6Gq691paAaty4UtX7t7UjfS9PW44xpkaMWx7qUrwXZ9QymuSV\nUsptTz9dqeqdGnaiWb1m7D+2n02HNtGlcZcABaZqOz1cr5RSpTAG/vAHePNNyMtzO5riROTE3vyy\ntGUuR6NqMk3ySilVinXr4OWX4cEHISrQxzxvucWWSujfxib5b3d+G4iIVIjQw/VKKVWKTz6xjxdd\nVPUBTfzWtvLDxvZp1QeAFbtWOB2NCiGa5JVSqhRffWUfL744CBt7/PFKL3Jum3OpG1WX1XtXszdz\nLy3iA9Udn6rN9HC9UkqVkJcHX39tnw8e7G4sZakbXfdEF7efb/7c3WBUjaVJXimlSvjuO9tffVJS\nAPur9zV6tC2VdEnnSwCYmzrX6YhUiNAkr5RSJXz2mX0cOjRIG0xKsqWShicNB+DLrV9SaAqdjkqF\nAD0nr5RSJXTpAuefD5deGqQNPvRQlRZrV78drRNas+voLlbuWknfNn0dDkzVdronr5RSJdx4oz0n\nf8klbkdSPhHh8lMvByBlW2YhCIgAACAASURBVIq7wYSJadOm0atXL+Li4khMTOTcc88lPz/f7bDK\npHvySinltpEj7eP771d60X5t+jFl1RS+2/2dw0GpkubNm8c999zD5MmT6d+/P8eOHWPjxo1EBbwj\nhaqruZEppZQLPv7YXmzXqxdEBOtYZ8+eVV703DbnAvBN2je1th97ecydmM0jplL1169fT9u2bbn4\n4otp1KgRAN26dQtEaI7Rw/VKKeVhDNx+O/TpAz/9FMQN33efLVVwauNTaRjbkD2Ze0jLSHM4MOVr\nzJgxxMbG0rhxY+Lj41m7dq3bIVVI9+SVUsrj119h1y5o1gx69HA7Gv+ICOe2OZdPN33Kip0raFe/\nndshVVpl96jdkJ+fz+9+9zt69+7N5MmTadCgAR06dHA7rArpnrxSSnksWGAfhw4N4qF6gBEjbKmi\nvq3tVfXLduhgNYEyZ84c1q1bx5QpU+jTpw+dO3cmMjLS7bAqpEleKaU8vEk+KF3Z+urXz5YqGtzR\ndsu3aOsipyJSJeTk5LB//37++9//sm3bNtatW8cbb7zBsWPH3A6tXHq4XimlgOxsWLzYPg9aJzhe\n995brcXPbnU2cdFx/PK/X9iZsZM2iW0cCkx5jRw5ktWrV/PQQw+xb98+GjRoQP/+/RkzZozboZVL\n9+SVUgpYssQm+l697Dn52iQmKoaLOl4EwIJNC1yOJjRFRUXx/PPPs2PHDnJycti3bx9z5sxxO6wK\naZJXSilg/35o0sSFvXiAK6+0pRqGdrKBf75FB6tRRfRwvVJKAaNGwe9+B1lZLmzcgaHuhnQcAtie\n72rr/fLKeZrklVLKIyIC6tVzYcP33FPtVXRp1IXGdRuz/9h+tqdvp32D9tWPS9V6fh2uF5FGIjJH\nRI6JyHYRub6C+nVE5FcR2elMmEopFThpaXD0qNtRVI+InBig5tud37ocjaop/D0n/yqQCzQHRgGT\nRKR7OfUnAP+rZmxKKRUUf/kLNGoEH3zgUgDDhtlSTd775VfsXFHtdQWSMTW/8xsn1IR2VpjkRaQe\nMAJ4yBiTaYxZCnwE3FBG/Q7AaOBpJwNVSqlAKCiAhQshP9/FXu6uuMKWahrQdgAAKdtTqr2uQImO\njibLlQsfgi8rK4vo6GhXY5CKfmmISC9gmTEmzmfevcBAY8xJ30oRmQ+8ARwGphtjSr1hU0TGAmMB\nmjZt2nvmzJlVbkRtl5mZSXx8vNthuEbbH77trwltX78+gXHjetOiRRbvvvstwbxezen2ZxVkcfnS\nyxER5g+YT2xkrGPrdkpkZCSJiYm0bt2ayMjIGj2CW1UZY8jNzWXXrl1kZGRQUFBQar1BgwatMsb0\nCWQs/ry78UBGiXnpQELJiiLyGyDSGDNHRJLLW6kxZiowFSApKckkJ5dbPaSlpKSg7U92OwzXhHP7\na0Lbl3l6gh0+vC6DBgU3lkC0v+fmnvyw5wc4BZI7O7tup2RkZLB//36OHj1KbGzN+yHihOjoaDp0\n6EBiYqKrcfiT5DOBklEmAsUuU/Ec1v8HcKkzoSmlVOD59lfvmiH29jcWVb9b2ks7X8oPe37g4w0f\nc0nnS6q9vkBITEwkMTGRlJQUevXq5XY4Ic2fC+82AFEi0sVnXg9gXYl6XYD2wBIR2QvMBlqKyF4R\naV/9UJVSylmHD9s9+agoR25Vr7rf/tYWB1x26mUAzN84v0Zc+KXcVeGevDHmmIjMBh4XkduAnsBw\noH+JqmuBtj7T/YFXgLPQK+2VUjXQt99CYSFccAHUr+9iIL//vWOrOrvV2TSNa8q2I9v49cCvdGva\nzbF1q9rH31vo7gTqAvuB94Bxxph1InK+iGQCGGPyjTF7vQU4BBR6pku/6kAppVx0ySV2/PiXXnI7\nEudERkRyaRd71nRe6jyXo1Fu8yvJG2MOGWOuMsbUM8a0M8a865m/xBhT6qWhxpiUsq6sV0qpmqJV\nKzjzTJeDSE62xSHDOtt77rUfexV69y4opZQfMjNtF7Y1oov3m292dHWDOgwCbKc4+YX5REXof/Xh\nSkehU0qFpXHjoFMn+PJLtyPBJnkHE32zes3o0qgLx/OOszxtuWPrVbWPJnmlVNjJzYV582DrVmhT\nE04q5uXZ4qArTrV9lX2y8RNH16tqF03ySqmw8+WXkJ4Op58Op57qdjTARRfZ4iDvxXdzU+c6ul5V\nu+iJGqVU2Jk1yz6OGOFuHCfcdpvjqzz/lPOJrxPP+gPr2X10N60SWjm+DVXz6Z68Uiqs5ObC7Nn2\n+bXXuhvLCaNH2+KgOpF16N/WdmeydMdSR9etag9N8kqpsDJ/vu3p7owzoHt5A2YH0/HjtjjsgnYX\nAPDl1ppwdaFygx6uV0qFlfnz7eNNN7kbRzGXeob8SElxdLUD2w8E4Ltd3zm6XlV7aJJXSoWV11+3\nh+nPPtvtSHyMGxeQ1fZo3gNBWLN/DcfzjhMXHVfxQiqk6OF6pVRYiYyEYcOgSRO3I/Hh4AA1vhJi\nEjir5VnkF+bzxZYvHF+/qvk0ySulwkJhIezc6XYUZUhPtyUAhicNB/RWunClSV4pFRbmz4f27WHC\nBLcjKcXw4bYEwBVJtlOczzd/rkPPhiE9J6+UCgv//CcUFNgBaWqcP/whYKs+s/mZNIlrQlpGGhsP\nbeTUxjWh9x8VLLonr5QKeatX2wvXExJgzBi3oynF1VfbEgAREsHgDoMBmL9hfkC2oWouTfJKqZD3\nr3/Zx1tvhcREd2Mp1YEDtgTIdd2vA2Dy95MDtg1VM2mSV0qFtL174b337JCyATwqXj3XXGNLgFyZ\ndCUNYhuw8dBG1uxbE7DtqJpHk7xSKqS99prtyvaqq6BjR7ejKcOf/2xLgERFRHFV16sAHZUu3GiS\nV0qFtAMH7F78H//odiTluOIKWwK5Cc/QszPWzQjodlTNokleKRXSXnsNNm+G8893O5Jy7N1rSwAN\n6zyM2KhYftz7I3szA7stVXNokldKhbwOHdyOoAIjR9oSQHWj6zKo/SAAFmxaENBtqZpDk7xSKiTN\nmgWffgq1ov+X++6zJcCGdR4GwKxfZwV8W6pm0CSvlAo5GRl2EJpLL4WvvnI7Gj9ccoktAXb1afZe\n/E83fkpaelrAt6fcp0leKRVy/vY3+9iwIQwa5G4sfklLsyXAWie25ppu11BgCpj1i+7NhwNN8kqp\nkHLgALz8sn0+aZK9sr7Gu+EGW4Lg6q52b/7DXz8MyvaUu7TveqVUSHnsMft48cUBGb01MLyHHoJg\nWJdhxETGsDxtOWnpabSt3zZo21bBp3vySqmQkZoKkydDRAQ8/7zb0VTCkCG2BEGD2AZcfurlGAxz\n1s8JyjaVezTJK6VCxj/+Afn5to/60093O5pK2LLFliD5TdffAGiSDwN6uF4pFTImTrT3xNfIkebK\nc+ut9jElJSibu+zUy4iKiOLr7V9z4PgBmsQ1Ccp2VfDpnrxSKmTUq2dPb7ds6XYklfTYY0UXEwRB\ng9gGDO4wmEJTyLzUeUHbrgo+TfJKqVrv3XfhyBG3o6iGgQNtCSI9ZB8eNMkrpWq1r7+G0aOhZ0/I\nznY7mipKTbUliIZ3HY4gfL75czJzM4O6bRU8muSVUrXWoUNw002269obboDYWLcjqqLbb7cliFrE\nt6Bf237kFOTw2abPgrptFTya5JVStVJBgR3TZds26N0bHn7Y7Yiq4amnbAkyPWQf+vxK8iLSSETm\niMgxEdkuIteXUW+CiKwVkaMislVEJjgbrlJKWQ88AAsXQtOmMHs2REe7HVE19O9vS5B5k/y81Hkc\nzTka9O2rwPN3T/5VIBdoDowCJolI91LqCXAj0BC4BLhbRAI7fqJSKuxMn27viY+MhJkzoV07tyOq\nprVrbQmyTo060a9NP47mHuWTjZ8Effsq8CpM8iJSDxgBPGSMyTTGLAU+Ak7qaNkY8w9jzA/GmHxj\nTCowFxjgdNBKqfB28KB9nDgRkpNdDcUZd99tiwt+2932/at92YcmMRUMtiwivYBlxpg4n3n3AgON\nMVeUs5wAPwBTjDGTS3l9LDAWoGnTpr1nzpxZtRaEgMzMTOLj490OwzXa/vBtf3Xa/t13DTnnnMMO\nRxRc3vYnrF8PwNGuXYMew/Zj27n5+5uJj4pndr/ZREcE77xHOH/3AQYNGrTKGNMnkNvwp8e7eCCj\nxLx0IKGC5R7FHin4T2kvGmOmAlMBkpKSTHJI/ByvmpSUFLT9yW6H4Zpwbn9l2r5+PcTFFR2aD4W3\n7ET7XW7MCzteYM3+NRxsdpCRpwfvDGs4f/eDxZ9z8plAYol5iUCZV2mIyN3Yc/OXGWNyqh6eUkrB\nmjU2D/bvD5s2uR1NAKxebYtLbu9tb9+btnqaazGowPAnyW8AokSki8+8HsC60iqLyK3AfcBgY8zO\n6oeolApnX38N558P+/ZB167QqpXbEQXA+PG2uOSabtcQFRHFwi0L2XZkm2txKOdVmOSNMceA2cDj\nIlJPRAYAw4G3S9YVkVHAU8BFxpjgDamklApJc+bA0KGQng4jRsD8+faQfcj5179scUnz+OZc1/06\nCk0h//mx1DOsqpby9xa6O4G6wH7gPWCcMWadiJwvIr79IT4JNAZWikimp5x00Z1SSpXHGJg0Ca65\nBnJyYNw4mDGjFvdoV5GePW1x0a097Uh4036aRkFhgauxKOf4NdSsMeYQcFUp85dgL8zzTndwLjSl\nVLhavx7uussm+8ceg4ceAhG3owqglSvt49lnuxbCoA6D6NiwI1sOb2HB5gVc2uVS12JRztFubZVS\nNc5pp9khY996y3ZXG9IJHmDCBFtcFCER/P6s3wMwddVUV2NRztEkr5RyXUYG3HEHfOYzTsrjj9tB\nZ8LCK6/Y4rJbet5CVEQU8zfMZ1fGLrfDUQ7QJK+Uco0x8N570K0bTJliO30rCMfTwaefbovLmsc3\n56quV1FgCvjPar0ALxRokldKuWLpUvjDH3px/fWwaxeccw783//Z/ujDzvLlttQAY88aC8C/f/i3\nXoAXAjTJK6WCatcuGDzY3vu+dm19mjWD11+3Oa4G7My644EHbKkBBnccTIcGHdievp3PN3/udjiq\nmjTJK6UCzneIjCZNYMMGSEyEG27YxoYNcNttYboH7zVlii01QLEL8H7QC/BqO03ySqmASU2F+++H\njh3hwAE7LyYGPvwQtm+HW2/dRv367sZYIyQl2VJD3NLLXoA3L3Ueu4/udjscVQ2a5JVSjtq6FV5+\n2Z5j79oVnnkGtm2DDz4oqnPOOdCggWsh1jyLF9tSQ7SIb8HwpOH2AjztAa9W0ySvlHJEYSH06GH3\n2v/wB9u/S0IC3HqrvcjujjvcjrAGe+QRW2qQsb3tBXiTvp9ERk7JgUhVbeFXj3dKKeVVWAg//wyf\nf24Hj5kzB6KjISICmjeH+vXthXXXXgtXXhmifc077c033Y7gJEM6DuGMZmewZv8anlj8BM8Nfc7t\nkFQVaJJXSpXr2DF7JHnlSlixwl4Fn+GzY/fddzBggH3+1lv2wroo/Z+lcjp2dDuCk0RIBC9e/CIX\nvX0RE7+byIQBE2hWr5nbYalK0j9FpRQAmZn2qvfUVHtx3NVX2/l79sBllxWv27YtDBkCF19c/La3\nFi2CF29IWbTIPg4Z4m4cJQzpOIRhnYfx6aZP+dOCPzH96uluh6QqSZO8UmEiP98+eveyP/zQDt26\nbRts2gQ7dxbVPeusoiTfqRNceqntT/7ss+G886B166CGHvqefNI+1rAkD/CvS/7FglcXMGPdDJ4d\n8iytE/XDr000yStVSxkDR49CdjY08xxFPXLEdoG+f//J5cABSEmBCy6wdVeuhGnTitZXpw507myv\niPcd9VQEPv44WK0KU2+/7XYEZTq18amMOG0EH/zyAY+mPMrrV77udkiqEjTJK+WSvDw4eBB27qzL\njz/ahJ2ZWfR43XX26nSAV1+FJUts/YMHbcLev9+OtX7ppUVJOC/PDstaGhF76N3rN7+xp4I7dLCl\nfXs9l+6atm3djqBcjyY/ytzUufz7x39zXffruKjTRW6HpPykf9KqRjDGlsLCohIRYfcuwSav9PSi\n1woKitdt3bqo7q5dcOhQ8de99evVgzPOsPUKC+HLL0+uk58Pubn20LT3eqhVq2DhQjs/N9cmV+9z\nY2Dy5KK2jB0LmzcXr5OdDVlZcPPNdnx0sLeVXXghQN9S35P+/e0hcrAXvM2YcXKdevWKJ+ZGjWzn\nM82anVwaN7ZXwXv17WuLqgG8w+9dcom7cZShW9NuPDrwUR748gFum3cba8etJSEmwe2wlB9CNsnf\neSds2VL6a5dfbke7Ali/Hu65p+z1TJpU9B/9Cy8UHwrTV9eutgMQsMli6NCirjxLPv7lL3bvC2D2\nbHj00Z4kJp5cNzLS7r15jR5tL4zyvu5b99pr4b777PSaNXDjjSfX8T6fNauoc62//c3GUFq9M88s\n6sAkP9+2saw2Pfus3fMEewj4wQeLJ1nfxH3wYFGbzjoLfvppIIWFJ7+nd95p92DBJtl+/U6u4/XT\nTzZeb5t8D0P76tvXJkywn9NF5eyQvPlm0Wf/zTc2eZYmKqp4kl+xwn4Gpdm3r+h5gwb2lrPIyCya\nNq1LQgLEx9vEnZhoH71uvx2GDbNJvHFjaNrUFt86YL8zTz1VdptUDfXMM/axhiZ5gAkDJvDhrx+y\nas8q7lt0H69e9qrbISk/hGySX7bM3stbGt+7VdLT7f2+ZTl6tOj5L78UXQRbku8tRcbYPcSy3Hhj\n0fM9e2DNmtK7/irZl/eaNWW3yXeP7PhxWL267O1nZxc9370bfv219HoleyTbvLnsdWZmFt/+7jJ6\nwowo0f2STf4C2MPJERFFxbf90dE2uZV8vbS6bdrYK75963rrdO9ePJbBg0+uFx1try4/5ZSiur17\n2x9nderY1+rUKf7cGBs/2B+G2dnF68TEQGysTdJevXrB3r2QkvItycnJZb+52IvdVAh7/323I6hQ\nVEQU/xn+H3pN6cXkVZO565y76Na0m9thqQqI8R05wiVJSUkmNTXV0XUuXVo88fhq27boP/sjR+Db\nb8teT79+nNjLXreu7OSVmFiUaH2TvPc/ft/HpCRo2dJO794Ns2b9SK9evUqt279/0TZ+/tke8vV9\n3fu8WTNo184+P3asaI/ft473eefOULeunbdrl/2hU3K7IjYpeROdMUVJvrQ4mzQpOn989Khdp28S\n9i2+fZXn5cHXX6dw4YXJJ9YXblJSUipM8qEqnNsOtbP94+aPY/KqyQzuMJgFoxcQGVH1kYVqY/ud\nJCKrjDF9ArmNkN2T93fPp0EDe6+vP7p3L74nWBYRu4foj1at4Mwz0zn//Irreg9HV6RePbuX6I/W\nrf27Hcr748AfCQlFCb8i0dH2x0C4JnilAJg3zz5ecYW7cfjh4YEP88EvH/DF1i+4Ze4tvHHlG0RH\nRle8oHKF9l2vlFJue+EFW2qBlgktmXHNDCIlkrd/fpsxH42hJhwRVqXTJK+UUm6bNcuWWmJwx8F8\nceMXREdE8/bPb/PAFw9ooq+hNMkrpZTbmjSxpRYZ2H4gky6bBMAzy57hrk/u0kRfA2mSV0opt82e\nbUstM+asMUz/zXQEYdL3k3h22bOa6GsYTfJKKeW2iRNtqYVGnTmKt37zFgD3f3E/g98azN7MvS5H\npbw0ySullNvmzrWllhp95mj+M/w/xEXH8dW2r7j0nUs10dcQmuSVUspt9esX70CiFrq5582su3Md\nLeJb8OPeH+n/Rn9W7V7ldlhhT5O8Ukq5bcaM0gcnqGXaN2jP97//nj6t+rD1yFb6v9mf99fW/N78\nQpkmeaWUctukSbaEgNaJrVlyyxLG9RlHbkEuo2aPYszcMXr43iWa5JVSym2ffGJLiIiNiuW1y17j\nkYGPUGgKeXP1m3Se2JmHvnyI9Ox0t8MLK5rklVLKbXFxtoSYR5MfZe24tVzW5TKO5R3jySVP0nFi\nR55b9hxZeVluhxcWNMkrpZTbpk+3JQR1b9ad+dfPZ8ktSzi/3fkcyjrEXxb9hVYvtmJm2ky2HC5j\nTHDlCE3ySinltn//25YQdl6781h882I+HfUpvVv25kj2ESZtmUSniZ3o+kpX/t8n/495qfN0D99h\nITsKnVJK1RoLF7odQVCICJd0voSLO13Me2vf49XFr7I6YzWpB1NJPZjKKytfoV50PS7pfAkDTxnI\nuW3O5fRmp1M3uq7boddafiV5EWkEvAEMBQ4A9xtj3i2lngDPALd5Zv0buM9oP4dKKVW26PAaqlVE\nuP6M62l1sBX9z+/Pd7u+44stXzBvwzxW7VnFh79+yIe/fghAhESQ1DiJni160r1pd1omtKRVQita\nxLegfkx96sfWJzEmkagI3Wctjb/vyqtALtAc6Al8LCI/GWPWlag3FrgK6AEYYCGwFZjsTLhKKRWC\npk2zjzff7GYUrqgTWYfz2p3Hee3O45HkR9h+ZDufbvqUJTuW8NPen1h/YD2/HviVXw/8Wu56YqNi\nia8TT3ydeOpF1yMuOu5Eia8TT2JMIrFRscRExhATFUN0RDRREVFER3oeI6LLfO6t5/t8QNsB2P3a\nmq3CJC8i9YARwOnGmExgqYh8BNwA3Fei+k3AC8aYnZ5lXwB+jyZ5pZQqWxgn+ZJOaXAKd/S5gzv6\n3AFAdn426/avY/Xe1Ww4uIE9mXvYdXQXB44fID07nfScdDJyMsjOzyY7P5sDxw8EPEZBKHykMODb\ncYI/e/KnAvnGmA0+834CBpZSt7vnNd963UtbqYiMxe75A+SIyFo/YglVTbCnQcKVtj982x/ObYeS\n7a8Fe4YOq5Wfv8EgjzryWSU5sZLy+JPk44GMEvPSgYQy6qaXqBcvIlLyvLwxZiowFUBEvjfG9PE7\n6hCj7df2h2v7w7ntoO3X9sv3gd6GP7fQZQKJJeYlAkf9qJsIZOqFd0oppVTw+ZPkNwBRItLFZ14P\noORFd3jm9fCjnlJKKaUCrMIkb4w5BswGHheReiIyABgOvF1K9beAP4lIaxFpBfwZmOZHHFP9Dzkk\nafvDWzi3P5zbDtp+bX+AiT9H0j33yb8JXAQcxN77/q6InA98aoyJ99QT4FmK3yf/Vz1cr5RSSgWf\nX0leKaWUUrWP9l2vlFJKhShN8koppVSIcjzJi0iKiGSLSKanpJZ4/XoR2S4ix0Tk/zzn+72vNRKR\nOZ7XtovI9U4tG0wi0sXzHkwvMT+k2y4i00Vkj4hkiMgGEbmtxOuDRWS9iBwXka9E5BSf12JE5E3P\nsntF5E9OLRsMnhje8Lz/R0VktYgMc6oNNb39njjuFpHvRSRHRKaV8npIt78yasrfbFWV91mHw+dc\n0d97jXoPjDGOFiAFuK2M17pj76+/ANtxzrvA+z6vvwfM8Lx2HrYzne7VXTbYBfgcWAJMD6e2e+KM\n8TzvCuwFenumm3jiuhaIBZ4DVvgs+7TnPWsInOZZ9pLqLhvEttcDHgXaY388X+75zNqHQ/s9cVyN\nHbtiEjCtxGsh3/5Kvlc14m/W6c86XD7n8v7ea9p7EIjGp1B2kn8KeNdnuhN24JsEz5uWC5zq8/rb\nwDPVXTbIH/5IYKbnC+Cb5EO+7SXehyRgD3CdZ3ossLzEH0kW0NUzvRsY6vP6E3h+yFRnWZffg5+B\nEeHWfuBJTk7yYdN+P96fGvk368RnHc6fs/fvvaa9B4E6J/+0iBwQkWUikuwzv1jf9saYzXi+7JTd\nR353B5YNChFJBB4HSjuEEtJt9xKR10TkOLAem+Q/8bxUsg3HgM1AdxFpCLSk7HEPqrOsK0SkOfaz\n8XYGFVbtL0W4t99XjfqbdVhYfs4l/t5r1HsQiCT/V6Aj0Bp7o/88Eenkea1k3/ZQ1A9+RX3kV2fZ\nYHkCeMN4RuErIdTbDoAx5k7Pts/HdqKU43mpojbAyeMe+Nv+8pYNOhGJBt4B/muMWe+ZHTbtL0O4\nt99XjfqbdVjYfc6l/L3XqPegUkle7EV1poyyFMAY860x5qgxJscY819gGXCpZxXl9YNfUR/51Vm2\n2ipqu4j0BIYA/yxjFbW27eDfZ+9ljCkwxiwF2gDj/GwDnDzugb/tL29ZR/jbfhGJwB56zQXu9llF\nWLS/HLW6/Q4Lyt+sS8Lqcy7j771GvQeVSvLGmGRjjJRRzitrMcA7Jl+xvu1FpCMQg+0fv6I+8quz\nbLX50fZk7EUXO0RkL3AvMEJEfnAgflfbDlX+7KOw1w+U1oZ6ntfWGWMOYw/tlzXuQXWWdYQ/7RcR\nAd4AmmPPxef5rCLk21+BWt1+hwXlb9YlYfM5l/P3XrPeA4cvPGgAXIy9KjAKGAUcw3OBCfbcQQb2\nUG49YDrFrxJ/H3vVaT1gACdfYV6lZYNRgDighU95HpgFNA31tntiaIa96DAeiPR8D44BV3peb+qJ\na4Tn+/Esxa8afQZYjL1qtCv2y3xJdZcN8nswGVgBxJfyWji0P8oT39PYvZtYICpc2l/J98r1v9lA\nfNbh9DmX9fde094DpxvdFFiJPXxwxPMGXFSizvXADmwCmAs08nmtEfB/ntd2ANc7tawLX4BH8bm6\nPtTb7vnsF3s+9wxgDfD7EnWGYC/Iy8LehdHe57UY7PgIGcA+4E9OLRuk9p+CPWqVjT2s5i2jwqH9\nPt95U6I8Gi7tr+R75frfbKA+63D4nCv6e69J74H2Xa+UUkqFKO3WVimllApRmuSVUkqpEKVJXiml\nlApRmuSVUkqpEKVJXimllApRmuSVUkqpEKVJXqkAK6c7WN+yzVN3moiUNvZB0InIo57Yopxcnx/1\nkj3bTXZiu0qFM0f+eJVS5epXYnoOdvSoR33m5aCUUg7TJK9UgBljVvhOi0gOcKDk/OoSkRhjjP5Y\nUEqdoIfrlaqBRKSXiCwRkeMislFE7ijx+s2eQ9oXiMgHInIE+Nbn9YEi8oWIHBWRYyKyQEROL7GO\ni0VkuYiki0imiKSKcH8OGgAAA0BJREFUyMOlhNNBRD721NkuIg97Rt/yXVeSiMwRkSMikiUiK0Tk\nEj/a2VRE3hWRDM+yb2HHwFBKOUCTvFI1TyLwLnYgouHY8SAmicigUuq+A2wFrgHuAxCRy4AvsH1p\nj8aOe5AALBGRtp46HYGPPMv+FrgSeBE7YEpJc4Avgauw/a0/BtzkfVFEWgFLsSNi3Q1chx3D4GMR\nGVZBW2cDlwMPeOLIB16uYBmllJ/0cL1SNU8CcKcx5isAEfkaO6rf74CvStSdZYz5S4l5LwGLjTHD\nvTNE5CtgC/BnYDxwFlAHGGeMyfBU+7KMeF4wxvzH83yRiFzoicU770/YUbH6GWM2ebb3CfAL8Hfg\n09JWKiIXAecBvzPGvO+ZvUBEPgXalBGLUqoSdE9eqZrnuDfBA3jOs28A2pVSd47vhGeM8k7AOyIS\n5S3AceAb4AJP1dVAHvC+iFwjIs3KiefjEtNrS8RyAXY4zE0+MRdgh1LtKSKJZay3H1AAfFhi/vul\n1FVKVYEmeaVqnsOlzMvBji9d0p4S095k/QY2ifuWy4HGAJ6EfDH2/4C3gb2e8+gDS9nGoQpiaVRK\nHAB7AcHu5ZemJXDYGJNXYv6+MuorpSpJD9crVbuVvO/8oOfxfmBRKfVzTyxojxZ8JSIxwADgcex5\n9PbGmAOViOEQ0KKU+S088ZX2owXsD4OGIhJdItE3r8S2lVLl0CSvVGhJBbYB3Y0xz/izgOd0wJci\nEg/MBToAlUnyi4Hxnh8H2wBEJBJ7Id2PPuf8S/oGiARGUPwQ/chKbFspVQ5N8kqFEGOMEZG7gLki\nUgeYiU3YzYH+wA5jzIueW/IuAD4B0oAm2L3/3dhz7pXxT+BmYKGIPAJkAHcCpwKXlRPrQhFZCkwR\nkSbARuwPg9PLWkYpVTl6Tl6pEGOM+QSbwOsB/wYWAP/AHj7/xlPtJ8/rTwOfA69gb6e70BiTVcnt\n7cZeJb8OmATMwp6nv8wY81kFi1+N/aHxNDADu+Nxd2W2r5QqmxhTYVfSSimllKqFdE9eKaWUClGa\n5JVSSqkQpUleKaWUClGa5JVSSqkQpUleKaWUClGa5JVSSqkQpUleKaWUClGa5JVSSqkQ9f8B25BA\nljv+gmYAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 576x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Cz1urmBqsJFK",
        "colab_type": "text"
      },
      "source": [
        "A curva PRC então pode ser graficada a a seguir"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Vb5nx1r_4fBm",
        "colab_type": "code",
        "outputId": "2cb14160-c446-443c-ec11-37ded1a0491c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 401
        }
      },
      "source": [
        "def plot_precision_vs_recall(precisions, recalls):\n",
        "    plt.plot(recalls, precisions, \"b-\", linewidth=2)\n",
        "    plt.xlabel(\"eficiência\", fontsize=16)\n",
        "    plt.ylabel(\"precisão\", fontsize=16)\n",
        "    plt.axis([0, 1, 0, 1])\n",
        "    plt.grid(True)\n",
        "\n",
        "plt.figure(figsize=(8, 6))\n",
        "plot_precision_vs_recall(precisions, recalls)\n",
        "plt.plot([recall_90_precision, recall_90_precision], [0., 0.9], \"r:\")\n",
        "plt.plot([0.0, recall_90_precision], [0.9, 0.9], \"r:\")\n",
        "plt.plot([recall_90_precision], [0.9], \"ro\")\n",
        "plt.show()"
      ],
      "execution_count": 26,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAgMAAAGACAYAAAAwF2G8AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 576x432 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bifaJ5IV5paH",
        "colab_type": "code",
        "outputId": "d06328cd-a59d-4916-a446-6002328d5a68",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68
        }
      },
      "source": [
        "y_train_pred_90 = (y_scores >= threshold_90_precision)\n",
        "print(precision_score(y_train_5,y_train_pred_90))\n",
        "print(accuracy_score(y_train_5,y_train_pred_90))\n",
        "rec_point = recall_score(y_train_5,y_train_pred_90)\n",
        "print(\"{:2.4e}\".format(rec_point))\n"
      ],
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "0.900159562343287\n",
            "0.9689166666666666\n",
            "7.3456e-01\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KBM-hmPBswm8",
        "colab_type": "text"
      },
      "source": [
        "A curva ROC relaciona outras métricas: o *false positive rate* e o *true positive rate* (recall)."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xqP_FrfM6yrT",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.metrics import roc_curve\n",
        "\n",
        "fpr, tpr, thresholds = roc_curve(y_train_5, y_scores)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SxOxnnWR7Oku",
        "colab_type": "code",
        "outputId": "f9296fb9-637f-47ea-c218-c73845e61194",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "fpr_point = fpr[np.argmax(thresholds<=threshold_90_precision)]\n",
        "print(\"{:2.4e}\".format(fpr_point))\n",
        "\n"
      ],
      "execution_count": 29,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "8.0368e-03\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cKosLKSXtkHi",
        "colab_type": "text"
      },
      "source": [
        "No gráfico o ponto de 90% de pureza tem 0.8% de taxa de falsos negativos."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qtTf1Nvj65nQ",
        "colab_type": "code",
        "outputId": "82a3368e-bbe4-46bd-9199-07e3853ff78b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 401
        }
      },
      "source": [
        "def plot_roc_curve(fpr, tpr, label=None):\n",
        "    plt.plot(fpr, tpr, linewidth=2, label=label)\n",
        "    plt.plot([0, 1], [0, 1], 'k--') # dashed diagonal\n",
        "    plt.axis([0, 1, 0, 1])                                    # Not shown in the book\n",
        "    plt.xlabel('FPR (1-efB)', fontsize=16) # Not shown\n",
        "    plt.ylabel('TPR (eficiência sinal)', fontsize=16)    # Not shown\n",
        "    plt.grid(True)                                            # Not shown\n",
        "\n",
        "plt.figure(figsize=(8, 6))                         # Not shown\n",
        "plot_roc_curve(fpr, tpr)\n",
        "plt.plot([fpr_point, fpr_point], [0., rec_point], \"r:\") # Not shown\n",
        "plt.plot([0.0, fpr_point], [rec_point, rec_point], \"r:\")  # Not shown\n",
        "plt.plot([fpr_point], [rec_point], \"ro\")               # Not shown\n",
        "                        # Not shown\n",
        "plt.show()"
      ],
      "execution_count": 30,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAgMAAAGACAYAAAAwF2G8AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\np2RAREQkzykZEBERyXNKBkRERPLc/wMPkl7ONs1U4gAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 576x432 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WKihKC9Jt7HP",
        "colab_type": "text"
      },
      "source": [
        "Um possível índice de performance é a área sob a curva ROC (AUC)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "t4I7u4LSDMnf",
        "colab_type": "code",
        "outputId": "664d60f9-f831-4a6f-ad52-5b486e63816d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "from sklearn.metrics import roc_auc_score\n",
        "\n",
        "roc_auc_score(y_train_5, y_scores)\n"
      ],
      "execution_count": 31,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.9741394824488451"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 31
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "dEJ6Be5YuFeI",
        "colab_type": "text"
      },
      "source": [
        "Agora podemos comparar com um modelo de **RandomForest**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WQufDA0yFDRA",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.ensemble import RandomForestClassifier\n",
        "forest_clf = RandomForestClassifier(n_estimators=100, random_state=42)\n",
        "y_probas_forest = cross_val_predict(forest_clf, X_train, y_train_5, cv=3,\n",
        "                                    method=\"predict_proba\")"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IoQWVGzkFO2g",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "y_scores_forest = y_probas_forest[:, 1] # score = proba of positive class\n",
        "fpr_forest, tpr_forest, thresholds_forest = roc_curve(y_train_5,y_scores_forest)\n",
        "prec_forest, rec_forest, threshs_frst = precision_recall_curve(y_train_5, y_scores_forest)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kP4S-4hNqyr2",
        "colab_type": "code",
        "outputId": "e0afaf57-2c9f-42a9-8036-7416169a5a5a",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "thr_frst_90 = threshs_frst[np.argmax(prec_forest>=0.9)]\n",
        "print(thr_frst_90)\n",
        "y_pred_frst_90 = (y_scores_forest >= thr_frst_90) \n",
        "rec_pnt_frst = rec_forest[np.argmax(y_pred_frst_90)]\n",
        "fpr_pnt_frst = fpr_forest[np.argmax(y_pred_frst_90)]\n",
        "print(rec_pnt_frst,fpr_pnt_frst)"
      ],
      "execution_count": 34,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "0.24\n",
            "0.9696800595238095 7.322788517867604e-05\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "z7lc_-B8FPtv",
        "colab_type": "code",
        "outputId": "3a88eec2-3380-4709-f34b-ddc0d34fbf66",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 401
        }
      },
      "source": [
        "plt.figure(figsize=(8, 6))\n",
        "plt.plot(fpr, tpr, \"b:\", linewidth=2, label=\"SGD\")\n",
        "plot_roc_curve(fpr_forest, tpr_forest, \"Random Forest\")\n",
        "plt.plot([fpr_point, fpr_point], [0., rec_point], \"g:\")\n",
        "plt.plot([0.0, fpr_point], [rec_point, rec_point], \"g:\")\n",
        "plt.plot([fpr_point], [rec_point], \"go\")\n",
        "plt.plot([fpr_pnt_frst, fpr_pnt_frst], [0., rec_pnt_frst], \"r:\")\n",
        "plt.plot([fpr_pnt_frst], [rec_pnt_frst], \"ro\")\n",
        "plt.grid(True)\n",
        "plt.legend(loc=\"lower right\", fontsize=16)\n",
        "plt.show()"
      ],
      "execution_count": 35,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAgMAAAGACAYAAAAwF2G8AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nKaWUKuU0GVBKKaVKuf8PWgU25Pjb+68AAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 576x432 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Vexd_YwsL-im",
        "colab_type": "code",
        "outputId": "e826ac7e-ba5d-40f8-c058-1e183601d1af",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "roc_auc_score(y_train_5, y_scores_forest)"
      ],
      "execution_count": 36,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.9984563317302746"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 36
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pAj0Me6oH7hj",
        "colab_type": "code",
        "outputId": "3b4db6d0-3ece-4ee5-83f2-1730d14aa559",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 292
        }
      },
      "source": [
        "plt.plot(recalls,precisions,\"b:\")\n",
        "plot_precision_vs_recall(prec_forest,rec_forest)\n",
        "plt.plot([0,rec_point],[0.9,0.9],\"r:\")\n",
        "plt.plot([rec_point],[0.9],\"ro\")\n",
        "plt.plot([0,rec_pnt_frst],[0.9,0.9],\"r:\")\n",
        "plt.plot([rec_pnt_frst],[0.9],\"ro\")\n",
        "plt.show()"
      ],
      "execution_count": 37,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZMAAAETCAYAAADzrOu5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n+//3cfkY0X5gFgAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QLXJVGahPIM3",
        "colab_type": "code",
        "outputId": "70988db0-cf96-4309-c6d2-bdf6f516673d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "from sklearn.metrics import f1_score\n",
        "\n",
        "print(f1_score(y_train_5, y_train_pred_90))\n",
        "print(f1_score(y_train_5, y_pred_frst_90))"
      ],
      "execution_count": 38,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "0.808972651848817\n",
            "0.9381452554089236\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "umkl6teHMHvA",
        "colab_type": "code",
        "outputId": "b6d196e0-9eec-4038-abd6-3c1f16f647e0",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 286
        }
      },
      "source": [
        "plt.plot(prec_forest,prec_forest*rec_forest,label=\"RandomForest\")\n",
        "plt.plot(precisions,precisions*recalls,\"b:\",label=\"SGD\")\n",
        "plt.grid(True) \n",
        "plt.xlabel(\"pureza\")\n",
        "plt.ylabel(\"eficiência * pureza\")\n",
        "plt.legend(loc=\"upper left\")\n",
        "plt.show()"
      ],
      "execution_count": 60,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYkAAAENCAYAAAD6/JlzAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0\ndHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAAgAElEQVR4nOzdd3hUVfrA8e9JhyQEEiD0Jh1Cl7pA\nUEFRERHFFUSxIGURlZ+IWFERBVddC4qogAsIrihNQBQlNJGmQEIJvQSBhBYy6Zmc3x9nQmJISIZM\nSXk/zzNP5t65c8+bS5h37qlKa40QQgiRFw93ByCEEKL4kiQhhBAiX5IkhBBC5EuShBBCiHxJkhBC\nCJEvL3cH4GiVK1fW9erVc3cYDpOYmIi/v7+7wygW5Fpkk2thyHXIVtRrsWPHjnNa6yq595e6JFGv\nXj22b9/u7jAcJiIigvDwcHeHUSzItcgm18KQ65CtqNdCKXU8r/1S3SSEECJfkiSEEELkS5KEEEKI\nfJW6Nom8pKenExMTQ0pKirtDsVtQUBD79u1zdxgu4+fnR61atfD29nZ3KEIIykiSiImJITAwkHr1\n6qGUcnc4dklISCAwMNDdYbiE1prz588TExND/fr13R2OEIIyUt2UkpJCSEhIiUsQZY1SipCQkBJ5\nxydEaVUmkgQgCaKEkH8nIYqXMlHdJIQQxVV8cjp7/orn+Pkk4hJSUUDr2hXp1rAynh7u/9IkScJF\nPD09CQsLIyMjg/r16zN37lwqVqxY5PMeO3aMO++8k6ioqCKfa9iwYaxbt46goCAAHn30UcaOHVvk\n8+YlIiICHx8funbt6pTzC1EcxSen8/Pes6RbMzlwNoF1B+I4EpeY57FNqwVyb/taVCjnTYe6lagX\n4o+HG5KGJAkXKVeuHDt37gTg4YcfZvr06bz44otujupq77zzDvfee6/d77NarXh6ehb6+IiICAIC\nAiRJiFJPa82Woxf47+ZjrNkXS1pGJgC+Xh50uSGEge1qEVYziBuqBlA10Jd0ayZr9sXy+vI9TF6R\n3bOxUdUAvni4A3VDXDsNiSQJN+jSpQu7d+8GwGKx0L9/fy5evEh6ejqTJ0+mf//+HDt2jL59+9Kp\nUye2bdtGzZo1Wbp0KeXKlWPHjh08+uijAPTp0+fKeVNSUhg1ahTbt2/Hy8uL9957j169ejFnzhyW\nLFlCYmIiBw8e5NlnnyUtLY25c+fi6+vLypUrCQ4OzjfeBQsWMGXKFLTW3HHHHUydOhWAgIAARowY\nwZo1a5g+fTrlypVj3LhxWCwWKleuzJw5c6hevToffvghM2bMwMvLi+bNm/P2228zY8YMPD09mTdv\nHh999BHdu3d34hUXwvWsmZqVkaeZuf4IkafiqVTem8Ed6zCgbU0qlvemaqAf5Xyu/mLl7enBXa1r\n0Kd5KMlpVs5ZUll3II5//xRN3w82MOeRjnSsn///V0crc0niteV72PvXZYees3mNCrzar0WhjrVa\nrfzyyy889thjgBkXsHjxYipUqMC5c+fo3Lkzd911FwAHDx7kiy++YM6cOQwaNIjvvvuOBx98kEce\neYSPP/6YHj16MH78+Cvnnj59OkopIiMj2b9/P3369OHAgQMAREVF8eeff5KSkkLDhg2ZOnUqf/75\nJ8888wz//e9/efrppwEYP348kydPBmDu3LmEhIQwYcIEduzYQaVKlejTpw9Llizh7rvvJjExkU6d\nOvHuu++Snp5Oz549Wbp0KVWqVOGbb77hxRdfZNasWbz99tscPXoUX19fLl26RMWKFRk5ciQBAQE8\n++yzDvt3EKI4yLBm8sPu03z060EOxyXSoLI/bw5oycB2tfDzLvzdtp+3J37enlTy96FRaCC3tazG\nQ7O28sjsrax+pge1KpV34m+Rrcz0bnK35ORk2rRpQ7Vq1Th79iy9e/cGzK3oCy+8QKtWrbjllls4\ndeoUZ8+eBaB+/fq0atUKgPbt23Ps2DEuXbrEpUuX6NGjBwBDhw69UsbGjRt58MEHAWjatCl169a9\nkiR69epFYGAgVapUISgoiH79+gEQFhbGsWPHrpzjnXfeYefOnezcuZOwsDC2bdtGeHg4VapUwcvL\niyFDhrB+/XrAtLMMHDgQgOjoaKKioujduzdt2rRh8uTJxMTEANCqVSuGDBnCvHnz8PIqc99LRBmR\nmJrBrI1H6flOBE9/sxNvTw+mD27HmnE9GdKprl0JIi+1KpXnv492JN2qmbHusIOiLliZ+x9b2G/8\njpbVJpGUlMStt97K9OnTGTt2LPPnzycuLo4dO3bg7e1NvXr1rowT8PX1vfJ+T09PkpOTr7v8nOfy\n8PC4su3h4UFGRsZ1ndPPz+9KO4TWmhYtWrB58+arjluxYgXr169n+fLlvPnmm0RGRl5XeUIUR4di\nE5j3+wm+2xFDQmoGN9arxKS7WnBz06oOb2iuVak8d7SqztI//2J8n6YElXf+zARyJ+Fi5cuX58MP\nP+Tdd98lIyOD+Ph4qlatire3N2vXruX48Txn672iYsWKVKxYkY0bNwIwf/78K6917979yvaBAwc4\nceIETZo0KVK8HTt2ZN26dZw7dw6r1cqCBQvo2bPnVcc1adKEuLi4K0kiPT2dPXv2kJmZycmTJ+nV\nqxdTp04lPj4ei8VCYGAgCQkJRYpNCHeKuZjE0wv/5Jb31vP1lhPc3Kwq343qyrcju9K7eajTeiIN\n794AS1oG87Zc+7PCUcrcnURx0LZtW1q1asWCBQsYMmQI/fr1IywsjA4dOtC0adMC3z979mweffRR\nlFJ/a7gePXo0o0aNIiwsDC8vL+bMmfO3O4jrUb16dd5++2169ep1peG6f//+Vx3n4+PDokWLGDt2\nLPHx8WRkZPD000/TuHFjHnzwQeLj49FaM3bsWCpWrEi/fv249957Wbp0qTRcixIlPimdT9cdZtam\noyhgVPgNPPaP+lQOKNr/tcJqXqMCDSr7szLyNCN6NMDL07nf9ZXW2qkFuFqHDh107kWH9u3bR7Nm\nzdwUUdGUpbmbsuT37yULzGSTa2G48jpYbG0On284QkJKBve0rcn/3dqEmhXLuaT8nJbt+ouxC/7k\n48FtubNVDcAhiw7t0Fp3yL1f7iSEEOIaUtKt/HfzMT6NOMzFpHR6Nw9lXO/GNKtewW0x3RFWnUnL\n9hARHXclSTiLJAkhhMhDZqZm6a5TvPNjNH/Fp9C9UWX+r08T2tQu+kwJReXpoWhevQLRZ5zfridJ\nQgghctl8+DxvrtxL1KnLtKxZgX8Pak3XGyq7O6y/aVotkLm/H0dr7dSJMSVJCCGEzc6Tl3j3p2g2\nHDxHjSA/3r+/Nf1b13TLnEkFqVmpHKkZmVxMSifY38dp5UiSEEKUeftOX+bdnw6wZt9Zgv19eOH2\npjzUpV6RB8A5k7+v+fi2pGRIkhBCCGc4dSmZd1dH8/2fpwj08+LZPo0Z1q0+Ab7F/6OxViXTq+rY\n+UTqhDhvio7ifyVKiTfffJOvv/4aT09PPDw8+Oyzz2jfvj2vvPIK3377Lf7+ZmbH++6778rssJ6e\nnrRo0QKr1YqXlxcPPfQQzzzzDB4eMgZSiKJISEnn04jDfLnxKGDGOozscYNLRjA7SsMqAQAcP58I\nVHFaOZIkXGDz5s388MMP/PHHH/j6+nLu3DnS0tJ46aWXOHPmDJGRkfj5+ZGQkMC777575X3lypVj\n06ZNBAYGEhsby+DBg7l8+TKvvfaaG38bIUquDGsmC7ae4D9rDnI+Mc2tYx2KKiuhXUhMd2o5kiRc\n4PTp01SuXPnK6OfKlSuTlJTE559/zrFjx/Dz8wMgMDCQSZMm5XmOqlWrMnPmTG688UYmTZoky3wK\nYacdxy/y8pIo9p6+TOcGwcy5vTlhtYLcHdZ18/XypEqgL39duv453QqjTNZbhIfDnDnmeXq62Z43\nz2wnJZntb74x2/HxZvv77832uXNme/lys33mTMHl9enTh5MnT9K4cWNGjx7NunXrOHToEHXq1LFr\nNHWDBg2wWq3ExsYW+j1ClHXnLak8t2gXAz/9jYtJaXwypB0Lhncu0QkiS93g8hw9l/fKdo5SJpOE\nqwUEBLBjxw5mzpxJlSpVuP/++4mIiPjbMbNnz6ZNmzbUrl2bkydPuidQIUoRa6Zm3u/HuenddXz/\nxylG9GzAmnE9uT2seqm5E69esRxnE1KcWoZLq5uUUsHAl0Af4BwwUWv9dR7H+QIfAAMAb2ATMFJr\nfcoRceT8fPb2/vt2+fJ/3w4K+vt25cp/365WrXBlenp6Eh4eTnh4OGFhYXz22WecOHHiytxMjzzy\nCI888ggtW7bEarXmeY4jR47g6elJ1apVC1eoEGXUzpOXeGVpFLtj4uncIJg3+rekUWjpmwOtfkh5\nVkaeJi4h1WlluPpOYjqQBoQCQ4BPlVJ5LfDwFNAFaAXUAC4CH7kqSEeLjo7m4MGDV7Z37txJkyZN\neOyxxxgzZsyV9SOsVitpaWl5niMuLo6RI0cyZsyYUvMtSAhHu5CYxvPf7WbAJ5s4HZ/CB/9sw4Lh\nnUtlggC4tWU1rJmajYfinFaGy+4klFL+wECgpdbaAmxUSi0DhgLP5zq8PrBaa33W9t5vgPdcFauj\nWSwWnnzySS5duoSXlxcNGzZk5syZBAUF8fLLL9OyZUsCAwMpV64cDz/8MDVqmAm7kpOT6dat25Uu\nsEOHDmXcuHFu/m2EKH6smZoFW0/wzupoLKkZPNatPk/d0ohAv5LTpfV63GDrBnvqYjKVnPSV32VT\nhSul2gKbtNblc+x7Fuipte6X69gOmOqm+4BLwBdArNb66XzO/QTwBEBoaGj7hQsX/u31oKAgGjZs\n6MDfxnWsVuuV1d/KikOHDhEfH3/VfovFQkBAgBsiKn7kWhgWi4UzGeWYuzeN45czaRrswdBmvtQM\nLDvNrU/+mki7ql7cVy+9SH8TvXr1cvtU4QHA5Vz74oG87gMPAieBU4AViATG5HdirfVMYCaY9SRy\nz6m+b9++ErsmQ1lcT8LPz4+2bdtetV/WUMgm1wIuJqbx1Oy1rI9JIbSCLx8+0Jx+rUpPo3RhNdiz\nCauvFwEByU75m3BlkrAAuSdgrwDkNdftdMAXCAESgeeAVUAnZwYohCj+tNYs2hHDW6v2cykpg+Hd\n6/PULY1LxFQazlAlwLljJVx5VQ8AXkqpRlrrrFbc1sCePI5tA7yotb4AoJT6CHhdKVVZa33uegp3\n9nS6wjFK20qJwrEOnk3gxSVRbD16gfZ1K3F3LU+G3tHc3WG5lZ+3BynpefeIdASXVdxprROB7zEf\n9v5KqW5Af2BuHodvAx5SSgUppbyB0cBf15sg/Pz8OH/+vHwAFXNaa86fP39lBLoQWZLTrEz9cT99\nP9hA9JkE3r4njG9HdKF2GWp7yI8GcOL3X1ffn40GZgGxwHlglNZ6j1KqO7BKa53V6vIs8CGmbcIH\niMKMmbgutWrVIiYmhrg453UTc5aUlJQy9aHp5+dHrVq13B2GKEZ+3X+WV5buIeZiMve2r8XEvk0J\nCfB1d1jFhwYPpbClC4dzaZKwVR/dncf+DZiG7azt85hxFA7h7e1N/fr1HXU6l4qIiMizEVeI0i72\ncgqvLd/LisjTNKoawDdPdKZTgxB3h1XseHgorJnOqyUpmy09QohiKzNT8832k0xZuY/UjEye7dOY\nJ3rcgI+XVC3lJbi8NxcS0zCVLo4nSUIIUWwcirXwwuJIth69QOcGwUwZEEaDKjIe5FqC/X2JT04n\nI9M5AwclSQgh3C4tI5MZ6w7z8a+HKOfjybSBrbivQy3pkVgIwQHmDsKSVgraJIQQIrcdxy/w/HeR\nHIy1cGer6rzarwVVAqVhurCq2JJEvCQJIURpcjklnWk/7mf+lhPUCCrHrGEduKlpqLvDKnGqVjC9\nHy+lSpIQQpQCWmt+jDrDq8v2cM6SyiNd6/N/fRrjX0ZHTBdVZX9z13VZkoQQoqQ7dSmZV5dGsWZf\nLC1qVODLh28sFSvEuZOft+n1lZ7pnPMXOkkopXyAF4EHgDqYxYCu0FqXralKhRCFZs3UzPntGO/+\nFI3W8NIdzRjWtR5entKttah8vcxHr9uTBPAGcD/wFvA+MB6oB/wTeNnhkQkhSoWoU/FM/D6SyFPx\n9GpShdf7t6R2cPmC3yiKBXuSxCDMEqI/KqX+DSzVWh9WSu0DegOfOSVCIUSJlJSWwfs/H+DLjUcJ\n9vfl48FtuaMUrS9dXGjbdBzOuqr2JIlQYK/tuQWoaHv+IzDVkUEJIUq2tdGxvLQ4ilOXkhncqQ4T\nbmtKULnSvUqcu2TNW1ocksQJzHrTJ4BDwK3ADsxa1M6bzFwIUWLEJaTyxg97WbbrLxpWDeDbkV24\nsV6wu8Mq1a70aXJSlrAnSSwGbgZ+xywtukApNRyoCbzjhNiEECWE1ppvt8fw5sp9JKdZeeaWxowM\nb3ClUVU4T9YSCG6/k9BaT8zxfJFS6iTQDTigtf7BGcEJIYq/I3FmvqXfj1ygY71gptwTRsOqMt+S\nqzh7lZzrHiehtd4CbHFgLEKIEiQtI5PP1h3mo7WH8PPy4O17whjUoTYeHtIwXZrYM07CCkQA92qt\nL+bYH4pZNU7uK4UoI3Ycv8jE73dz4KyFO1pV59V+zakaWHYWxypOsqZQzygG4yQUUAHYopS6U2t9\nINdrQohSLiElnWk/RjNvy3GqV/Djy4c7cHMzmW/JnXw8s5KE+6fl0Jg1qV8CfldKDdJar8nxmhCi\nFPtpzxleWbqHswkpDOtaj2f7NJH5looBXy8PPBSkWp1zfnvGxCsgQ2s9GjPCeplSarRzwhJCFBex\nl1MYNW8HT8zdQcXy3iwe3Y1X+7UocQkiLQ2mTIHYWNi+HV57LXuMQUmmlKJSeR8s6cXjTsI80Xq6\nUmo/8D+gh8OjEkK4XWamZuG2k7y1yiwjOv7WJjzRowHexXy+pbQ0eO89uPdeqF4dAgJgyRJITYUX\nX4QtW2DZMnPsfffBxInw8ceQkQG7d0P//u6N/3oElfMmMT3FKee2t03iCq31L0qpzsAyx4YkhHC3\nQ7EWXvg+kq3HLtClQQhT7gmjfmV/d4eVr5gYqF0bEhPBw8N88Fut0LGjef3CBRg6FFq1glmzYNcu\nkyjmzTM/a9eGlSvh6FFYtw569oTvv4evvoKlS2HfPliwAKpUgTFj3Pu75kUp59X52zNO4qqvD1rr\ng0qptpgpO4QQJVxJWEb06FHzrX/LFnjzzXZERcF335nXTp6EevXghx+gWzeoWPHvVUq7dpmfN91k\nHidOQN26MHy4Od7DA86cMcdUqQJbt5rnvr7w+uvm+a23QuPG8OST5o7Fq5jUurl9MB2AUsoPuBO4\nAfhMa30JM+L64jXfKIQo9nJ2a3X3MqKZmZCebj6cd+6Etm3h1ClYtQoefxyaN4f27SE6OpBTp2Ds\nWFO9VKOG+VZ9xx2FK6dOHRgxwjx/4IHs/YMGmZ9//ZW979tvzfknTzbbd9wB3rbpqNzdtuHM4gtd\nuaiUagjsB2YAbwJZE7KMAqY5PjQhhCtcTknnpSWR3DvjNywpGcwa1oGPB7dzeYJIToaoKPN8yhTw\n84Nt2+CPP8y+DRtMdVLr1qaN4Ysv4Jdf1lGnjkkMNWuan85y773QtaupgtIaOnUy+7/8EipUgLff\ndl7ZBXJilrDnTuI/wE+YpHApx/5lwGxHBiWEcD6tNauizjDJtozosK71+L8+TQhwYa+l1FRTbdS9\nu/nQjYyEQ4dMlRFAw4Zw443wyCPZCWDs2Oz3u7MWLKsqy2KBxx4zSe7QIahUCUJCXB9Pcahu6gp0\n1lpbc9VPZs0OK4QoIU5dSuaVJVH8st8sI/rFwx1oVatiwW90gKyqGa2halW4fBkOH4aXX4bff4cb\nbjCPBx/Mfk8xahK5SkAA7N9v2jcaNTL7XF395Mzi7P3KkNeE8HWAeAfEIoRwsgxrJnN+O8Z7Px9w\n6TKiEREQHQ333GMSw8iR8OmnMGcOxMVBgwbmcd99Tg3DaZo0MY3ZDRrA4sWmx1TfvtltFs6mtXZa\nIrUnSfwEjAMes21rpVQF4DVghaMDE0I4VmRMPBMX7ybq1GWXLiOalga9epnnnTubsQt9+5rtAQOc\nXrzLKGXuiCZMgGnToE8fWL3a3VEVnT1JYhywVikVDfgB3wANgbOYpU2FEMVQYmoG7/50gDm/HSUk\nwDXLiK5eDbfdZhqY+/eHRYvMN+2wsL/3GCqNpkwx4y4eftgkjthY053WmTyUwklTN9k1TuIvpVQb\n4AGgHaZn1ExgvtZaVqYTohj6Zd9ZXlm6h1OXkhnSqQ7POXEZ0fXrTQPz3r1m8BlkN0APHOiUIosl\nT08z4C6rq+zhw6Y7bw0nttx6e3pgdWeSUEp5A/OAF7TWs4BZzglHCOEIsQkpTFq2h5WRZ2gcGsB3\no7rQvq7jlxH96y/zoRgaasYbHDli9j/9tHmUZS+9ZKrZPDxM99w//jDjPZzBx8uD9DTnnLtQrVVa\n63SgDzLbqxDFmtaab7ad4JZ317FmXyzjb23CD092d0qCsFigRQt45RWzvXkzJCWZAXDC6NbN/AwI\ngFq1skd8O5q3p8JaDKYK/x64B/i3UyIRQhTJsXOJTPw+ks1HztOxfjBv3xNGgyqOXUZ0/Xozr9GM\nGebO4aOPoEsX81pF1/SgLXE6doSEBHPHlZlppg6pVcuxZTizfcmeJHECeEkp1R3YDiTmfFFr/Z4j\nAxNCFE6GNZMvNx7lvZ8P4OPpwZQBYfzzRsctI5qaatoY2rQxYxrA9FCCv49lENf2+eewcSNcuuT4\n0eHaiQMz7EkSwzBzNLWyPXLSQKGShFIqGPgSU311Dpiotf46n2PbYUZ6t8MkpSla6w/siFmIUi3q\nVDwTvtvNnr8u07t5KG/0b0m1IMcuI9qypRlJnJICd97p/nmKSqpHHzWTCYaFmW1HXsdiMZhOa13f\nQWVOB9IwM8e2AVYopXZprffkPEgpVRn4EXgGWAT4AA6+SROiZEqzat5atY8vNhylUnkfPhnSjr4t\nqzmk2sFqNfMULVlivvVOn24Gwvn4OCDwMq57d1PV9PbbcPGimcLDUZSTJuZw6SS3Sil/YCDQUmtt\nATYqpZYBQ4Hncx0+DlittZ5v204F9rksWCGKqd8On+PlTcmcTTrCoA61ePH25gSVL3q31l9/NXMn\nPfWUWUMBoFw5MyisT58in15gEu3x46Z9AkxC9nDAYHdn3t2pwtZlKaU+vNbrWuux13rddo62wCat\ndfkc+54Femqt++U69lcgErgRM2hvC/AvrfWJPM77BPAEQGhoaPuFCxcW/AuVEBaLhYAAxzY+llRl\n/Vokpmu+iU5jfUwGlf00j4aVo3mIp0PObbF40q9fd8qXz2DFio1kZCi8vIp/vVJJ/Zvo1Suc2bO3\nkpmpaNAgseA3FOD1zcn4KisTOl//tejVq9cOrXWH3PvtuZMIy7XtDTQFPIE/C3mOAOByrn3xQGAe\nx9bCtEX0xiSLacACoFvuA7XWMzED++jQoYMODw8vZDjFX0REBKXp9ymKsnwtVkWeZtKyPZy3ZDCi\nRwPa+Z7h1pt7Femca9ZA796mx9Kdd8KmTdC2rRflyoU7JmgXKKl/E1rDRx91ZOzY7FHpRTF9/29Y\nLsc75VoU+kZHa90r1+MfmA/ylZi1rgvDAlTIta8CkJDHscnAYq31Nq11CmaOqK5KqaDCxixESXf2\ncgoj5m5n1Pw/qBLgy9J//YOJtzfD1/P66p+PHMke5JZV5ZGZaX527Wqql4RrDBpkJgB0xOe6v68X\nKRlFP09eitQmobVOUUpNwTQwzyjEWw4AXkqpRlrrg7Z9rYE9eRy7m7832hf/e18hHCQzU7Nw20ne\nWrWP1IxMnrutCcO7N8C7iLO13n23aXeYMsWMBs7MLN7TcJdmoaFmxPpDD5lBdseOXf+5/H29SHbS\nvByOmB+4MqYaqUBa60TMoLzXlVL+SqluQH9gbh6HzwYGKKXa2KYFeRnYqLWWaclFqXYkzsIDn//O\nC4sjaVGjAquf7sHo8IbXlSBSUkwS6NjRbK9da+ZWKm9rFZQE4V4VKphpxUeMyL6jux4BPsXgTkIp\nNS73LqA6MART5VRYozFzP8UC54FRWus9tkF6q7TWAQBa61+VUi9gpiEvD2wEBttRjhAlSro1k5nr\nj/DBLwfx9fJg6sAwBnWobXe31sRE+M9/zCysqalmX9aazyEh7lk1TeTNx8e0T+zaBT/9ZP7Nroev\ntwfpxWBajidzbWcCcZhv/G8V9iRa6wvA3Xns30CuOxKt9afAp3bEKESJtOvkJSZ8t5v9ZxLo27Ia\nr93VgqoVrm9Q3IEDZnK50FB4/HEZ/FYStGkD999//UkCisHypQ4cTCeEsElKy+C9nw4wa9NRqgT6\n8tnQ9tzaoprd55kyxQx627HDzDS6Zw80b+6EgIVTzJyZPRK7uLmuNgmlVIBSquR1ThaiGFl/II4+\n76/ni41H+WfHOvw8rqddCeLixezqpAsXTCNo1pABSRAly/DhZuK/Qde5fJsz7xbtShJKqaeVUicw\nYxvilVInlVLPKGdOQShEKXMxMY1x/9vJQ7O24uPpwTdPdGbKgDAq+BV+1PSJE+UJDs7uzjptmvmg\nKIHjyoTNoEFFW7XP7dVNSqlpmFHN7wCbbbu7AK9gGrCfc3h0QpQiWmuW7fqL15fvJT45nSdvasi/\nejXEz7two6Z37jS9k555BmrWTCI8HEaPNq85YmoH4V4XLkBUlJktdvhwd0eTzZ6G68eBx7XWi3Ls\n+9W25vVnSJIQIl+nLiXz0uJI1kbH0bp2ReYPDKNptdzjSq/t8cdNm8OgQWYg3Nq1TgpWuEWlSrBq\nFbz1lvm3Li71M/YOptudzz75HiNEHqyZmnm/H2faj/vJ1PDync0Z1rUenoVY6yE5GZ580izq89hj\n8OmnpsdSzZpw8GCBbxclUEAAjBwJGRlmNHZhaSeONbYnSfwX+BfwVK79o8h7MJwQZdrBswlM+G43\nf5y4RPdGlZkyIIzaweULfJ/Vau4U/Pzg8GGob+tXeOONTg5YuN0LL5h/86zZYuvUsePNTrrzsCdJ\n+AKDlVK3Ar/b9nUCagDzc3piIwAAACAASURBVM4SW5gZYYUordIyMvkk4hDT1x7C39eL9wa1ZkDb\nmoUaFLdwITzwgKlKCg+HX36R9oayZssW87O4LAdrT5JoCvxhe17X9vOM7dEsx3EydEeUWX+cuMjz\n3+3mwFkLd7WuwSv9mlM5wPea77l8GX7/3azZkLVedNbyoJIgyp7Bg80XhMC85sZ2A3sG0xVtXmIh\nSrHE1AzeWR3NV5uPUa2CH18+3IGbm4UW6r2VK0N6uqlmqltXRkgLs+jT6NGQlla4tgln/s24dGU6\nIUqjiOhYXlwcxalLyTzUpS7jb21C4DXGPKSmwoQJppdS166maunoUblrENlWrTI/4+PNl4jCcPs4\nCSHE311ITOONH/ay+M9T3FDFn0Uju9ChXnCB7zt3Dj74AGJiTJLo1s08hMiybJmZFbY4fHGQJCGE\nnbIGxb22fC+Xk9MZe1NDRhcwKO6110yD5IoVpgvr/v3QuLELgxYlzvDhZlr3zZsLPtaZJEkIYYfc\ng+KmXmNQnNbmrqFiRZg/34xtiIuDqlWhSRMXBy5KnDNnoFOnwh3rzGYsSRJCFEJmpmbeluNMXVW4\nQXFpadC0qZmLJzoatm41C8wUh+oDUTKsWGHvO5zTKlGoJKGUega4qLWeo5QaAlTVWr/vlIiEKGYO\nxVp4/rvdbD9+scBBcQkJpuuij0/22gB16+Z5qBDXFBUFw4bB3LnQrFmBhztNYe8kZgI/K6XWYFaW\n6+28kIQoHtKtmXy27jAf/nKIcj6e/Pu+1gxsl/+guJUr4Z//NG0PzZrBJ5+4OGBRqpw6Zebqcvcc\nTgUmCaXUQ7ane4CtwA/AvUoptNb/dWZwQrjL7phLPLfIrBR3R1h1Jt3VgiqBVw+Ks1pNu0NoqFlH\neuBA8Pd3Q8Ci1Ln1VvjxR1NtWRB3j5NQeTwvJvMTCuFYyWlW3l9zgC82HKFKoC8zh7anzzUWAurf\n3ySJ334z/dlnz3ZhsKLUO38e3ngDXn654GOddcdRYJLQWn+llPIHRmLmaloIPK21TnJOSEK4x2+H\nzjFxcSTHzyfxQMfaPN+3GUHl/j4oTmuzTOhDD5mG6EcfNY3U7q4SEKWP1jB2rGmXKMTRToujsG0S\nTwAztdYnlFKfACMAabgWpUJ8cjpvrdzHwm0nqRtSnq+Hd6LrDXkPc92zx0zfvWEDfPMN3HOPi4MV\nZcr+/YX/AuLWEdc5ezJprec5KRYhXG71njO8vCSKc5ZURvRowNO3NKacz98HxR0/DsuXw5gx0LIl\nfPEFDB3qpoBFmaEUJCXBf/4DU6deew6nYrPGtRClRWxCCqPn72DE3B2EBPiy9F//YOLtza5KEGDa\nGSZMyF5/+LHHTBdXIZzt8cfh/ffNOBt3kcF0okzRWrNoRwyTV+wjOd3K+Fub8ESPBnh7Zn9fysyE\nefOgRg245RYYP960PdSo4cbARZm0apX5kuLn574YJEmIMuPkhSReWBzJhoPnuLFeJd66pxUNqwZc\ndZzVClOmQLt2Jkn4+0u3VuEenp7w1VdmHqd9+/LvDuvuLrBClGjWTM2c347x79XReCh4o38LhnSq\ni0eOKTX274dWrWD3bvMf8Zdfshf+EcKdspY0LWgpU5kqXIjrEH3GrDO98+QlejWpwuQBYdSsWO6q\n4ywWs/DPrl0mSdSs6YZghchD376FO85ZNxN2JQmlVC/gAaAO8LemO631TQ6MS4giSc2w8snaw3wS\ncYgAXy/+c38b+repcWVKDa2hdm3Tg+TkSejQwfQkKXd1/hDCrQ4dMtPKL1qUf5drHy8PrJnOSROF\nThJKqWHADGAxEA4sBRoD9QHpFiuKjT9OXGTCot0cjLXQv00NXrmzOSG2daYzM80x6elmbhwwg+F8\nfCRBiOLJw8O0j13r79PXy4P0TOeUb8+dxLPAGK31F0qpBGCi1vqIUupjwOKc8IQovMTUDP79UzRz\nfjPrTM8a1oGbmmavMx0fb9Z26NjRTMIna0mLkqBBA9MF9lrTzPt6F48k0QBYY3ueCmR1C/kYiACe\nd1xYQthnw8E4Jn4fSczFZIZ2rstzt2WvM52UBOXLZ/dQeuiha5xIiGLohRfMUrd33ZX3675enlg1\nZFgz8fJ07PA3e5LEeSDQ9vwU0BLYDYQAcqMu3OJSUhqTV+xj0Y4YGlT2538jutCxfvY609OmmYFw\nMTGmMVruHkRJo7XpBuvpmX+SyFr8KiNT45X/KrrXxZ4ksQHoA0QC/wM+VEr1Bm4GfnZsWEJcm9aa\nVVFneGXpHi4mpTE6/AbG3twIP29PrFZTtRQcbJYJbdBAVoQTJZdScPq0WYTIajXJIq9jnMWe/zpj\ngAW2528B72DuIv4HPF7YkyilgpVSi5VSiUqp40qpwQUc76OU2qeUirEjVlGKnb2cwoi5Oxg9/w+q\nBfmybEw3nrutKX7e5n9PtWoQEmKO7d/f9DGXMQ+iJDt0CO6+Gz77zPVlF/pOQmt9IcfzTGDqdZY5\nHUgDQoE2wAql1C6t9Z58jh8PxJFd1SXKKK01C7ee4M2V+0jLyOT5vk15/B/18fL0IC7OrOegFIwc\nCZUqmdt0mcJblAZ+fpCY6J5lTK+ZJJRSwVnJQSkVfK1jcyaRa5zPHxgItNRaW4CNSqllwFDyaPhW\nStUHHgTGAZ8XdH5Reh0/n8i0bSnsuxBJp/rBvD2wFfUrm5boXbvgH/8ws7Pef79ZpEWI0qRWLVPl\n5A4F3UnEKaWqa61jgXPkPahP2fYXprmkMZChtT6QY98uoGc+x38EvAAkF+LcohTKsGYye9Mx3v05\nGqUzmTIgjH/eWBulFKdOmcboli3NbJnt27s7WiGcIy0Nfv0VWrd2fdVpQUniJiDrDqGXA8oLAC7n\n2hdPHlVJSqkBgKfWerFSKvxaJ1VKPYFZGInQ0FAiIiIcEGrxYLFYStXvY4+TCZnMikrlaHwmbat6\nMrCupkbyEdavP8KMGQ346adqzJ27BX9/K/37mx5MMWWk5aos/13kVFauQ1ycL4MGdaFv39M891z0\nVa8fOZIGwPr16/HxdGwd6zWThNZ6XV7Pi8ACVMi1rwKQkHOHrVpqGnB7YU6qtZ4JzATo0KGDDg8P\nL3KgxUVERASl6fcpjNQMK9PXHuaTzYcIKufNx4NbcEdYdX74YROtWv2D4GAzKK5TJ+jdu3uZXNuh\nLP5d5KWsXAetYcUK6NmzOv7+V99K7OUQHIimR48eVzpwOIo903LcB6RprZfm2t8f8NZaLyrEaQ4A\nXkqpRlrrg7Z9rYHcjdaNgHrABttcOz5AkFLqDNBZa32ssHGLkuXPExeZ8N1uDpy1MKBtTV6+sznB\n/j6cOAF33fUPhg+HmTOhTRvzEKIsUArCw+H33+EmF8+SZ08X2ElASh77E22vFUhrnQh8D7yulPJX\nSnUD+gNzcx0aBdTG9H5qg+lie9b2/KQdMYsSIjnNyuQf9jLw099ISMlg1rAOvH9/G+JjzW2CmYTv\nAqNGuTlQIdzE3x9Gj3Z9ufZOy3F1ZRgcsr1WWKOBWUAsZhT3KK31HqVUd2CV1jpAa50BnMl6g1Lq\nApCptT6T5xlFibb58Hme/343x88nMbhTHSb2bUqgnzdLl5pZLyMioHt3eOed3bRtG+7ucIVwi8hI\nOHfO9eXakyQuYqqBjuXa35hcbQrXYusqe3ce+zeQPR9U7tcigFqFLUOUDAkp6by1aj9fbzlB3ZDy\nLBjemY71Qjh1CgLrQO/e8PLLpkeHEGVdo0bg62tmMnblDAL2FLUUeF8p1Thrh1KqCfAesMTRgYnS\nbe3+WPq8v56FW08wvHt9fnyqB11uCOH++6FPHzOVd/nyMGkSVMjd1UGIMuiXX8y6Eh995Npy7bmT\nmACsAvYqpbKGdVQHtmJGRQtRoAuJaby+fA9Ldv5F49AAPn2wG5VVRZRtmuORI+HiRfCSNROF+Jvw\ncDNg9PZC9fl0HHum5bgMdLNN6pfVr+RP4BetZW5NcW1aa1ZEnubVpXuIT07nqZsb8a9eDYk760GT\nJmam1pdfhltucXekQhRP5cvDY4+5vly7v69prX9GZn0Vdjh7OYWXl0Tx096ztKoVxPzhnQjIqICP\nlxkx/cYbZvIyIUT+LBbT7XvsWPNwFXvXuO6EmRq8KrnaM7TWLgxblARaa77dHsMbK/aSlpHJC7c3\n5dFu9XlwiAfffJO9xsMzz7g7UiGKP6vVzHCctXiWq9gzmO5ZzCjoQ8Bf/H0eJ6luEn9z8kISLyyO\nZMPBc3SsH8yLvVtRycsfL08YNcr8wWdN5y2EKFhQEHz3nZnt2JXsuZN4Chirtf7YWcGIki8zU/PV\n5mO8szoaBbxxd0se6FCH1q0V1avDmjXQs6d5CCHss2QJnD0Lr7ziujLtSRIVgJXOCkSUfIdiLUz4\nbjc7jl+kZ+MqDG7YCu8kP7y84M03zXTHQojrt3WrWUSruCaJBcBtwCdOikWUUOnWTGauP8IHvxyk\nvI8n7w1qTbWUmnTrpujYEbZskYZpIRzhyy9NF3FXLqhlT5I4Cbxmm29pN5Ce80Wt9XuODEyUDFGn\n4pnw3W72/HWZvi2q8ViblnQI8yUzE/79b7POgxDCMeLi4NVXYdAgM27CFexJEo9jpvruanvkpDEj\nr0UZkZJu5aNfDzJj3RGC/X2Y8WA7HuldnRlnICEBAgLg//7P3VEKUbocPAizZ8PAga4r057BdPWd\nGYgoOXYcv8Bzi3ZzOC6R22+ox8R+jaldzZvHHzfrPLi6i54QZUXXrpCUZKqbXOW6Jj9QSoUCcVrr\nTAfHI4qxxNQM3lkdzVebj1EjqBz/uasTA7pVZl84rF0ra0sL4QobN5rxRQ3smXu7CAo9wZ9Sylsp\nNU0plQCcwiwKhFJqqlLKDbOcC1facDCOW/+znq82H+Puhg356Zke9OtUmdtvhw8+cHd0QpQd/fvD\ney6s3L9mklBKeSilXrJtvgr0Ax4EUnMcthUY5pTohNvFJ6fz3KJdDP1yKz5eHjxeoycfj2xC9B4v\nPD3NkoqtWrk7SiHKjuXLYcwY15WXb3WTUqopMIfssRGDgUe01uuUUjmrmaIwa0qIUmb1njO8vCSK\nc5Y0hrZpzIsDG5CS6EnyWTO3vRDC9Tp1MtOFN23qmvKudScxELNy3Ju27erA8TyO8+I62zZE8RSX\nkMq/vv6DEXN3UDnAlxt29ebHdxvh7eFJxYoweTIEBro7SiHKpgYNYNw40x3WFa714f4O8BqwFugB\n7LH9PJbruEHADmcEJ1xLa82Snad4bfleLAnwbN8mjOjZgBV1PYiPd+1qWEKIvG3ebB5BQa4pL98k\nobVOAyYqpTrYdr0GzFNK1QY8gftsVVKDgTucHqlwqr8uJfPi4kjWRsfRxD+Us1+3o0KYB96eMlpa\niOKkZk24917XlVfgd0Ot9Xbbz+WYu4Y+QCamIbsR0E9rvcaZQQrnyczUzPv9OH3eX89vBy7xyp3N\n+WFie+7oaxYDEkIUL1u3mpmUXTVWwq62BK31amC1k2IRLnbsXCITvtvNlqMXqB7bhOOrG3D3eA+8\nvcwyiUKI4ue112DlStP13MfH+eVJg3MZZM3UzNp4lHd/jsYLT6YODKOJd23eSVVYre6OTghxLStW\nQHp6MZngTyl1GWigtT5nG0SX7w2O1rqCo4MTjnfgbALjF+1m54lLePzahRtbVuD+G82fwfz5bg5O\nCFEoc+fCxx/DH384v6yC7iSeBBJyPJcV6EqotIxMPo04zEe/HqRCOW8+GtyWrd6VCApSLp12WAhR\ndEeOQH0XzaZ3zSShtf4qx/M5To9GOEVkTDzjF+0icq8V6689mTXPmx6tfbirtbsjE0Jcj8mTXVeW\nPXM39VRKXbXopG1/D8eGJRwhJd3K26v2c/cnm7iYlMaHw1pQLcCf9EQXtHYJIZwmNhY+/NDMCOts\n9jRcvw+8nsf+CsAkoL0jAhKOse3YBSYs2k3U7/6EXujM6h8CqVjem3v+lKolIUq6F180PRD79XN+\ntZM9Y2ibALvy2B9le00UA4mpGby6NIpBn20mzZrJkObN8LwQjDXZG5AEIURp8NFHcPQo1K7t/LLs\nuZNIxszfdDTX/ppAmsMiEtdtw8E4nvtfFPt/rs4d4S357Oma+Hl5oV8BL+nsLESp4ecH9eq5pix7\n7iRWA1OVUpWydiilgoG3kAF2bhWflM74b8103r5eHlQ61YjKl+ri72um85YEIUTp8sMPsGCBa8qy\n5+PjWWA9cEwptdu2rxUQC9zv6MBE4azec4bn50ezf24YfQeE8vUbVUgZYWZrFUKUTjNmwOnT8MAD\nzi/LnjWuTyulWgNDgDa23V8BX2utXdDGLnI6Z0nl1WV7WLH7NFUttUmJCaZXI/DzBj9JEEKUaosW\ngcXimrLsnbspCfjcSbGIQtBas2zXXzw/6yiXYvx5cUxjRvS8gbgXoEYNd0cnhHAFPz/zcIWCpuW4\nB1iutU63Pc+X1vp7h0YmrnI6PpmXFkfxy/5Y0jd2Qp8JZvgCM523JAghyo6334YePaBrV+eXVdCd\nxCKgGqbdYdE1jtOYNSYKZGvs/hIz5fg5YKLW+us8jhsPPAzUtR33idb6ncKUUdporVm47SQvTj+D\nR3ACL93TjNueDMHLU+Hr6+7ohBCulJICEyfCm28WgyShtfbI63kRTcd0mQ3FtG2sUErt0lrvyXWc\nAh4CdgM3AD8ppU5qrRc6KI4SITYpk8Gfb2FjZAJ/fX0zDz1q5fHu3u4OSwjhJn5+kJoKGRmuKe+a\nH/xKqSNKqRDb81eUUuWLUphSyh+zdvbLWmuL1nojsAwYmvtYrfU0rfUfWusMrXU0sBToVpTySxJr\npubz9Uf4v6/LEXUqnmlDm7BqhWL6+5IghCjrfHygfJE+jQuvoOqm6kB54DxmJboZQFF6MjUGMrTW\nB3Ls2wVcNSdUTkopBXQHPsvn9SeAJwBCQ0OJiIgoQojud8qSyazIVHZG1OLCT10Y859tVE86Aj5H\n2LrV3dG5j8ViKfH/to4i18Ioi9fBYvFkxIgOPPbYUW66KRaAI0fMeOb169fj4+nYaRUKShJ/ArOU\nUhsx1T/PKqXy7Hiltc5rXqfcAoDLufbFA4EFvG8S5q5ndj5lzwRmAnTo0EGHh4cXIpTiJ92aycc/\nH+bD1TFUrOLJJ5OC2dluL+OfvBEPR1X2lWARERGU1H9bR5NrYZTF6xAbCz17Qnh4c8LDmwOwl0Nw\nIJoePXrg512o5uFCKyhJPAJMBu7GNE73A/KqCdPkPflfbhbMhIA5VSB7zYqrKKXGYNomumutUwtR\nRokUdSqe8Yt2E/F+MwJVTbbs8qRqBV9CUw/i4dHc3eEJIYqJqlXh66u6+jhPQQ3X0cB9AEqpTKCn\n1jq2COUdALyUUo201gdt+1oDuRutsZX5KPA80ENrHVOEcoutlHQrU5cd5qvthwgJ8GHic550qFee\nqrLOnxCiGLBnxHWRKzy01olKqe+B15VSj2N6N/UHrurIpZQaAkwBemmtjxS17OJo+7ELPDnzANs+\nbEPvh/1Y+HJ1gspLw7QQIn9TpsDs2RAdjUuqoe0qQinVVym1Qim1TylV27bvcaXUzXacZjRQDjP2\nYgEwSmu9RynVPVd7x2QgBNimlLLYHjPsibe4SkzN4JXFe7jvs82ogCTuuFMz+dE6kiCEEAVq1Ahu\nucU1CQLsuJOwfbOfAXwB3ARkfaJ5As8BvxTmPFrrC5g2jtz7N2AatrO2XbSCq2ttPHiOxyed4fDK\n+jzzseLVexvj7yvTtAohCue++8zDVezJRc8Bw7XWz/D3xuvfyZ7wT+QjPjmd5xbt4sEvtxBYNYUu\n7b155qbmkiCEEHb57Tf4pVBfyR3Dnk+oRsDmPPbn1WNJ5LA66gzD/pVMSoY/E166gadvaeTwbmpC\niLLh/ffh2DHYts015dmTJP7CDIY7nmt/D+CwwyIqRc7bpvP+YfdpPFPb0btRMBNu85UlRIUQ123M\nGNe1R4B9SWIm8KGtVxJAbaVUd2AaZrCbyOH8efjXU57sDElk3F2NeeL1UPx8ZEScEKJoel5zfgrH\ns6cL7DSlVBDwM+AHrAVSgX9rrac7Kb4SKyUF1vzoxZRpXXjiZml3EEI4xj33QOPGZrpwV7B30aEX\nlVJvAs0xjd57tdYuWh+p+Dt0CPr3h1mzoFMnOHoUAgMlQQghHCc0FEJCXFee3Z9gttXptjshlhJv\n8WLYuzd7Ct/AgmakEkIIO336qWvLk0ryItq6FTbb+nw98wycOgXdysyE5kIIV5s9G5o3h40bXVOe\n1IUUgdUKQ4eapUPXrgUvL1lGVAjhXKdPw4ABptrJFSRJXIe1a6F7d5MUvv8eatd2d0RCiLLihRdc\nW55UN9npt9/gpptgzhyz3aIFVJChhEIIF9m0ybR/uookiULQGg7Y1tLr0gXmz4eHHnJvTEKIsumL\nL2DcONeVJ0miECZMgI4dzYpQSsHgwWaNWSGEcLXnn4f//td15UmSyEd6OiQmmufDh8O0aVC5sntj\nEkKIRo3M/E1BQaaWw9mk4ToPaWlmMFznzqZPcqNG5iGEEO7m4WESxOjRpoell5M/xSVJ5GC1gqen\nqUq6/37TF1kIIYqb2bPh5Emz+FBEhHPLkuommz/+MPOh7N5ttp9/Hu66y70xCSFEfkaONNXiziZ3\nEjaffmraHDIyCj5WCCHcbflyiI93fjmSJGxmzDBVTUIIURJ4eEClSi4ox/lFlAySIIQQJcWmTaZr\n/oULzi9LkoQQQpQwP/4IO3dCuXLOL0uqm4QQooS56y545BHXJAm5kxBCiBImOBieesp0088a9Oss\nkiSEEKKECQ42j2++AW9v55Yl1U1CCFHCVKoEX30Fr75qOt1U7uG8suROQgghSqgjR+DwYeeWIXcS\nQghRQs2da35+EuG8MuROQgghSiBXzAALkiSEEKJEGjgQRo1yfjlS3SSEECWM1tCkiVnW4PffnVuW\n3EkIIUQJoxS89RacOQMDBji3LLmTEEKIEmrsWOjdG5KcWIbcSQghRAmTng6ZmWYFzWHDnFuWJAkh\nhChBtIY774RHH3VNDyeXJwmlVLBSarFSKlEpdVwpNTif45RSaqpS6rztMVUppVwdrxBCuNvmzfDe\ne+a5UtCunVm61BWfiO5ok5gOpAGhQBtghVJql9Z6T67jngDuBloDGvgZOArMcGGsQghxlaQkiIkx\nK1nWqGE+rA8cMLOy1qpl9u/cCVWqQN26cPkybNhgJuRr0MC8d8kSuPVWaNQI/vzTrI45frxZRvl/\n/4Onn4b1683r69bBxInQuTOEhZlG6+Rk2L/flOdMLr2TUEr5AwOBl7XWFq31RmAZMDSPwx8G3tVa\nx2itTwHvAsNcFqwQQuRj82bTBbVFC/jwQ/jtN+jY0XyAf/aZWRSod29o0wZmz4YtW+DBB80dwPz5\nJoG89BLceCN8+y2cPWsSw403muTRqJFpb2jfHlauhDFjzDm6dYM1a0wMkZHQrJlJJM6ktKuG7QFK\nqbbAJq11+Rz7ngV6aq375To2Huijtd5i2+4ArNVaB+Zx3icwdx6Ehoa2X7hwoRN/C9eyWCwEBAS4\nO4xiQa5FNrkWhruuw4ULPuzYUYlz53y48caLhISkEhUVxNmzfrRrd5GQkDQOHQrgxIlytG17icqV\nUzlzphzR0YGEhV2ievUUEhO92Lo1mBYtLlOzZjJJSZ5s2hRCixaXqVEjBYvFk99/D6Fly3iqVUsl\nIcEcHxYWT9WqqcTHe7FjRzCtWl3iuDWJjSdS+Fd7f7w8rq8OqlevXju01h1y73d1kugOfKu1rpZj\n33BgiNY6PNexVqCF1nq/bbsRcADw0NcIukOHDnr79u3OCN8tIiIiCA8Pd3cYxYJci2xyLQy5DtmK\nei2UUnkmCVc3XFuACrn2VQASCnFsBcByrQQhhBDCsVydJA4AXra7giytgdyN1tj2tS7EcUIIIZzE\npUlCa50IfA+8rpTyV0p1A/oDc/M4/L/AOKVUTaVUDeD/gDkuC1YIIYRbBtONBsoBscACYJTWeo9S\nqrtSypLjuM+A5UAkEAWssO0TQgjhIi4fJ6G1voAZ/5B7/wYgIMe2Bp6zPYQQQriBTMshhBAiX5Ik\nhBBC5EuShBBCiHy5dDCdKyil4oDj7o7DgSoD59wdRDEh1yKbXAtDrkO2ol6LulrrKrl3lrokUdoo\npbbnNQqyLJJrkU2uhSHXIZuzroVUNwkhhMiXJAkhhBD5kiRR/M10dwDFiFyLbHItDLkO2ZxyLaRN\nQgghRL7kTkIIIUS+JEkIIYTIlyQJIYQQ+ZIk4WZKqWCl1GKlVKJS6rhSanA+x41XSkUppRKUUkeV\nUuNdHauzFfZa5DjeRym1TykV46oYXcWea6GUaqeUWq+UsiilziqlnnJlrM5mx/8RX6XUDNs1uKCU\nWq6UqunqeJ1FKTVGKbVdKZWqlJpTwLHPKKXOKKUuK6VmKaV8r7dcSRLuNx1IA0KBIcCnSqkWeRyn\ngIeASsBtwBil1D9dFqVrFPZaZBkPxLkiMDco1LVQSlUGfsRMox8CNAR+cmGcrlDYv4ungC5AK6AG\ncBH4yFVBusBfwGRg1rUOUkrdCjwP3AzUBRoAr11vodK7yY2UUv6YP+SWWusDtn1zgVNa6+cLeO+H\nmH+/J50fqfPZey2UUvWBlcA44HOtdS1XxutM9lwLpdQUoLbWeqjrI3U+O6/Fp0CC1vo52/YdwHta\n6yYuDtuplFKTgVpa62H5vP41cExr/YJt+2Zgvta62vWUJ3cS7tUYyMj647fZBVzr2zNKKQV0p3Qt\n52rvtfgIeAFIdnZgbmDPtegMXFBK/aaUirVVsdRxSZSuYc+1+BLoppSqoZQqj7nrWOWCGIubFphr\nlGUXEKqUCrmek0mScK8A4HKuffFAYAHvm4T5t5vthJjcpdDXQik1APDUWi92RWBuYM/fRS3gYUxV\nSx3gKGbFx9LCnmtxvx7wIgAABLZJREFUEDgJnLK9pxnwulOjK54CMNcoS9bzgj5X8iRJwr0sQIVc\n+yoACfm9QSk1BtM2cYfWOtWJsblaoa6FrfphGjDWRXG5gz1/F8nAYq31Nq11CqbuuatSKsjJMbqK\nPddiOuCLaZvxB76nbN5J5L5mWc/z/Vy5FkkS7nUA8FJKNcqxrzX5VCMppR7F1iCltS5tPXoKey0a\nAfWADUqpM5gPguq2nhz1XBCnK9jzd7EbyNmwWNoaGe25Fm2AOVrrC7YvUB8BHW2N+2XJHsw1ytIa\nOKu1Pn9dZ9Nay8OND2AhpnrAH+iGuTVskcdxQ4AzQDN3x+zOa4FZl71ajsc9mF4f1TBVUG7/PVz8\nd3ETpmG3DeANvA9scHf8broWs4HvgCDbtXgB08Dt9t/BQdfBC/AD3gLm2p575XHcbbbPiuZAReBX\n4O3rLtfdv3hZfwDBwBIgETgBDLbt7w5Ychx3FEjH3EpmPWa4O353XItc7wkHYtwduzuvBTAKUw9/\nEViO6e3k9t/B1dcCU800H4gFLgEbgY7ujt+B12ES5k4x52MSpi3KAtTJcew44CymbWY24Hu95UoX\nWCGEEPmSNgkhhBD5kiQhhBAiX5IkhBBC5EuShBBCiHxJkhBCCJEvSRJCCCHyJUlCCCFEviRJCCGE\nyJckCSGcTCnl4+4YhLhekiSEyEEpFWFbAvMDpdRF2+MdpZSH7fVjSqln83jPxzm2jymlJtmWjbyE\nmSoCpVRNpdTCHOddkXPiOqWUzuuR4/W3lVLRSqlkWxnTlFJ+Tr8o4v/bu3fQKKIwiuP/46OIjaUv\nROIzYKlggpDY2YhYKKiptDD4QFTUQlGQIIovUqXTLo2popUSiKlU8FUIBisrIygiSsTKz+Kb6Lib\nidlASMDzg4GdmTuXmWL323sH7vmvuUiY1eskvxttQBdwCDjRYB+ngBFgM3CuCMEZAn4AHUXfo8Bg\ncQ5gWWlbCTwHhkt9jgEHyZyEI8Be4HyD92XWkAWzfQNmc9AocDxyYbMRSevJH/1bDfQxHBHXxneK\nZd4FHCj6RVIXuRjdDuBuRHwote8lV/DcPn4sIrpL/b8roktPAxcafD6zKXORMKv3JP5e+fIx0C2p\nNvxmMs9q9jcBzcC3TJ/9bRGwpnxA0lFgP9AWpQwASbvJEc1aMn1sfrGZzRgXCbPG/CRHBGULJ2g3\nVrM/D3hFThHV+jz+oQitvwHsiog3peOtZK7CJeAkuRT2zqKt2YxxkTCrt0WSSqOJVuB9RHyV9JF8\nZwBA8eK4BXj5jz5fAPuATxHxZaIGxUvsfuBsRDyoOb2VDNDpLrVf1chDmU2HX1yb1VsO9EjaUEzx\nnCET3yBTvjolbZO0EbjD1P5s9ZEhMAOSOiQ1S2qXdFPSOklNwD1gEOiXtHR8K65/C6yQ1ClptaTD\nZNExm1EeSZjV6yPn+p+S6V+3+VMkrpAZ2wNkGthlsqhMKiK+S2oHrpKjhcVk7OoQmSi3hByRtAB7\nai5XRNyXdB3oAZqAh8BFoHe6D2k2FU6mMyuR9Ah4HRHHZvtezOYCTzeZmVklFwkzM6vk6SYzM6vk\nkYSZmVVykTAzs0ouEmZmVslFwszMKrlImJlZpV9qppwRHKD8ZgAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MHsC6EpbfeV_",
        "colab_type": "code",
        "outputId": "51dc393d-682b-49b2-e249-9ae31a2c61fa",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 268
        }
      },
      "source": [
        "y_scores_only_signals = y_scores[y_train_5]\n",
        "y_scores_only_backg = y_scores[~y_train_5]\n",
        "#h = np.histogram(y_scores_only_signals,bins=100,range(0,1000))\n",
        "plt.hist(y_scores_only_signals,bins=100,range=(-80000,20000),histtype=\"step\")\n",
        "plt.hist(y_scores_only_backg,bins=100,range=(-80000,20000),histtype=\"step\")\n",
        "plt.yscale(\"log\")\n",
        "plt.show()"
      ],
      "execution_count": 40,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX0AAAD7CAYAAACG50QgAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-yChCIpGil0G",
        "colab_type": "code",
        "outputId": "0c73753f-c870-484f-8e09-b8bb8d532445",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 268
        }
      },
      "source": [
        "y_scores_only_signals_frst = y_scores_forest[y_train_5]\n",
        "y_scores_only_backg_frst = y_scores_forest[~y_train_5]\n",
        "#h = np.histogram(y_scores_only_signals,bins=100,range(0,1000))\n",
        "plt.hist(y_scores_only_signals_frst,bins=101,range=(0.,1.),histtype=\"step\")\n",
        "plt.hist(y_scores_only_backg_frst,bins=101,range=(0.,1.),histtype=\"step\")\n",
        "plt.yscale(\"log\")\n",
        "plt.show()"
      ],
      "execution_count": 41,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAAD7CAYAAABkO19ZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8Ur4r7Ma_fle",
        "colab_type": "text"
      },
      "source": [
        "### Amostra \"reduzida\" (só 3 números)\n",
        "\n",
        "Aqui a prevalencia do \"sinal\"  (número 5) aumenta de 1/10 a 1/3. "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UhsSpHgySDKl",
        "colab_type": "code",
        "outputId": "e319227d-4937-4339-e5c4-17a97ac27b95",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "indx = [i for i,y in enumerate(y_train) if y in [\"3\",\"5\",\"8\"]]\n",
        "X_train_reduced = X_train[ indx]\n",
        "y_train_5_reduced = y_train_5[indx]\n",
        "print(X_train.shape)\n",
        "print(X_train_reduced.shape)\n"
      ],
      "execution_count": 42,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(60000, 784)\n",
            "(17339, 784)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "x3vv1dgWSZjl",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "forest_redx_clf = RandomForestClassifier(n_estimators=100, random_state=42)\n",
        "y_probas_forest_reduced = cross_val_predict(forest_redx_clf, X_train_reduced, y_train_5_reduced, cv=3,\n",
        "                                    method=\"predict_proba\")"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RX4XpLs1UC88",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "y_scores_frst_redx = y_probas_forest_reduced[:, 1] # score = proba of positive class\n",
        "fpr_frst_rdx, tpr_frst_rdx, thresholds_frst_redx = roc_curve(y_train_5_reduced,y_scores_frst_redx)\n",
        "prec_frst_rdx, rec_frst_rdx, threshs_frst_redx = precision_recall_curve(y_train_5_reduced, y_scores_frst_redx)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Qu7LTXByUh5o",
        "colab_type": "code",
        "outputId": "06559c8d-703c-45dc-e636-0258c0a859da",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 670
        }
      },
      "source": [
        "plt.figure(figsize=(8, 6))\n",
        "plot_roc_curve(fpr_forest, tpr_forest, \"Random Forest\")\n",
        "plt.plot(fpr_frst_rdx, tpr_frst_rdx, \"r:\", linewidth=2, label=\"RandomForestReduced\")\n",
        "\n",
        "plt.grid(True)\n",
        "plt.legend(loc=\"lower right\", fontsize=16)\n",
        "plt.show()\n",
        "\n",
        "plt.plot(rec_frst_rdx,prec_frst_rdx,\"r\",linewidth=2,label=\"RandomForestReduced\")\n",
        "plt.plot(rec_forest,prec_forest,label=\"Random Forest\")\n",
        "plt.xlabel(\"eficiência\")\n",
        "plt.ylabel(\"precisão\")\n",
        "plt.grid(True)\n",
        "plt.legend(loc=\"lower right\", fontsize=16)\n",
        "plt.show()"
      ],
      "execution_count": 61,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAgMAAAGACAYAAAAwF2G8AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nA0IIIUQ1979NsgOsFFj+AgAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 576x432 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYkAAAENCAYAAAD6/JlzAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nz+3vDOu4AAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4vsL9GCqAQe4",
        "colab_type": "text"
      },
      "source": [
        "Nos gráficos acima podemos ver que na amostra reduzida a curva ROC não muda, mas a curva PRC mostra uma melhoria."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BDgpm4KfffmC",
        "colab_type": "text"
      },
      "source": [
        "![texto alternativo](https://github.com/clemencia/ML4PPGF_UERJ/blob/master/pics/Thresh_vary_valassi.png?raw=true)\n",
        "\n",
        "![texto](https://github.com/clemencia/ML4PPGF_UERJ/blob/master/pics/ROC_PRC_prevalence.png?raw=true)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Bbp8FqW7GW6L",
        "colab_type": "text"
      },
      "source": [
        "![texto](https://github.com/clemencia/ML4PPGF_UERJ/blob/master/pics/TableDomains.png?raw=true)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "fc_j_ziEGY8f",
        "colab_type": "text"
      },
      "source": [
        "Podemos testar para o modelo MLP\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BrryudQ2vbkn",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 80
        },
        "outputId": "9cb76668-284f-49f3-8518-f387fd0e1851"
      },
      "source": [
        "from keras.models import Sequential\n",
        "from keras.layers import Dense\n",
        "from keras.optimizers import SGD\n",
        "from keras.datasets import mnist\n",
        "from keras.utils import np_utils\n",
        "\n",
        "# Trainign batch size \n",
        "batch_size = 64\n",
        "# Number of training epochs\n",
        "epochs = 20\n",
        "# Fraction of the training data to be used as validation\n",
        "val_split = 0.3\n",
        "\n",
        "# Learning rate\n",
        "lr=0.1\n",
        "\n",
        "X_train_mlp = X_train.astype('float32')\n",
        "X_test_mlp = X_test.astype('float32')\n",
        "X_train_mlp /= 255\n",
        "X_test_mlp /= 255\n"
      ],
      "execution_count": 46,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Using TensorFlow backend.\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              "<p style=\"color: red;\">\n",
              "The default version of TensorFlow in Colab will soon switch to TensorFlow 2.x.<br>\n",
              "We recommend you <a href=\"https://www.tensorflow.org/guide/migrate\" target=\"_blank\">upgrade</a> now \n",
              "or ensure your notebook will continue to use TensorFlow 1.x via the <code>%tensorflow_version 1.x</code> magic:\n",
              "<a href=\"https://colab.research.google.com/notebooks/tensorflow_version.ipynb\" target=\"_blank\">more info</a>.</p>\n"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YFIMKN89wqYe",
        "colab_type": "code",
        "outputId": "e5a289ef-bc4a-40ec-86d4-2da4a9357be7",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 445
        }
      },
      "source": [
        "# Multilayer Perceptron model\n",
        "model = Sequential()\n",
        "model.add(Dense(input_dim=784, activation=\"relu\", units=100, kernel_initializer=\"normal\"))\n",
        "model.add(Dense(activation=\"relu\", units=30, kernel_initializer=\"normal\")) # after first layer no need to specify input_dim\n",
        "model.add(Dense(activation=\"relu\", units=10, kernel_initializer=\"normal\")) # after first layer no need to specify input_dim\n",
        "model.add(Dense(activation=\"sigmoid\", units=1, kernel_initializer=\"normal\"))\n",
        "model.compile(optimizer=SGD(lr=lr), loss='mean_squared_error', metrics=['accuracy'])\n",
        "model.summary()"
      ],
      "execution_count": 47,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:66: The name tf.get_default_graph is deprecated. Please use tf.compat.v1.get_default_graph instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:541: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:4409: The name tf.random_normal is deprecated. Please use tf.random.normal instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/optimizers.py:793: The name tf.train.Optimizer is deprecated. Please use tf.compat.v1.train.Optimizer instead.\n",
            "\n",
            "Model: \"sequential_1\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "dense_1 (Dense)              (None, 100)               78500     \n",
            "_________________________________________________________________\n",
            "dense_2 (Dense)              (None, 30)                3030      \n",
            "_________________________________________________________________\n",
            "dense_3 (Dense)              (None, 10)                310       \n",
            "_________________________________________________________________\n",
            "dense_4 (Dense)              (None, 1)                 11        \n",
            "=================================================================\n",
            "Total params: 81,851\n",
            "Trainable params: 81,851\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qYlQGitLxDXW",
        "colab_type": "code",
        "outputId": "4ad8f90f-0ec1-4102-d6f9-b17725838108",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "history = model.fit(X_train_mlp, y_train_5, validation_split=val_split, epochs=epochs, batch_size=batch_size, verbose=1)"
      ],
      "execution_count": 48,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:1033: The name tf.assign_add is deprecated. Please use tf.compat.v1.assign_add instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:1020: The name tf.assign is deprecated. Please use tf.compat.v1.assign instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:3005: The name tf.Session is deprecated. Please use tf.compat.v1.Session instead.\n",
            "\n",
            "Train on 42000 samples, validate on 18000 samples\n",
            "Epoch 1/20\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:190: The name tf.get_default_session is deprecated. Please use tf.compat.v1.get_default_session instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:197: The name tf.ConfigProto is deprecated. Please use tf.compat.v1.ConfigProto instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:207: The name tf.global_variables is deprecated. Please use tf.compat.v1.global_variables instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:216: The name tf.is_variable_initialized is deprecated. Please use tf.compat.v1.is_variable_initialized instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:223: The name tf.variables_initializer is deprecated. Please use tf.compat.v1.variables_initializer instead.\n",
            "\n",
            "42000/42000 [==============================] - 3s 77us/step - loss: 0.0949 - acc: 0.9096 - val_loss: 0.0825 - val_acc: 0.9094\n",
            "Epoch 2/20\n",
            "42000/42000 [==============================] - 2s 51us/step - loss: 0.0809 - acc: 0.9108 - val_loss: 0.0817 - val_acc: 0.9094\n",
            "Epoch 3/20\n",
            "42000/42000 [==============================] - 2s 54us/step - loss: 0.0796 - acc: 0.9108 - val_loss: 0.0790 - val_acc: 0.9094\n",
            "Epoch 4/20\n",
            "42000/42000 [==============================] - 2s 51us/step - loss: 0.0736 - acc: 0.9108 - val_loss: 0.0692 - val_acc: 0.9094\n",
            "Epoch 5/20\n",
            "42000/42000 [==============================] - 2s 53us/step - loss: 0.0519 - acc: 0.9257 - val_loss: 0.0301 - val_acc: 0.9628\n",
            "Epoch 6/20\n",
            "42000/42000 [==============================] - 2s 52us/step - loss: 0.0220 - acc: 0.9736 - val_loss: 0.0180 - val_acc: 0.9783\n",
            "Epoch 7/20\n",
            "42000/42000 [==============================] - 2s 53us/step - loss: 0.0143 - acc: 0.9824 - val_loss: 0.0135 - val_acc: 0.9828\n",
            "Epoch 8/20\n",
            "42000/42000 [==============================] - 2s 52us/step - loss: 0.0110 - acc: 0.9870 - val_loss: 0.0121 - val_acc: 0.9851\n",
            "Epoch 9/20\n",
            "42000/42000 [==============================] - 2s 53us/step - loss: 0.0088 - acc: 0.9894 - val_loss: 0.0121 - val_acc: 0.9844\n",
            "Epoch 10/20\n",
            "42000/42000 [==============================] - 2s 50us/step - loss: 0.0075 - acc: 0.9912 - val_loss: 0.0092 - val_acc: 0.9887\n",
            "Epoch 11/20\n",
            "42000/42000 [==============================] - 2s 50us/step - loss: 0.0064 - acc: 0.9924 - val_loss: 0.0083 - val_acc: 0.9894\n",
            "Epoch 12/20\n",
            "42000/42000 [==============================] - 2s 48us/step - loss: 0.0057 - acc: 0.9932 - val_loss: 0.0082 - val_acc: 0.9899\n",
            "Epoch 13/20\n",
            "42000/42000 [==============================] - 2s 49us/step - loss: 0.0050 - acc: 0.9946 - val_loss: 0.0195 - val_acc: 0.9741\n",
            "Epoch 14/20\n",
            "42000/42000 [==============================] - 2s 49us/step - loss: 0.0046 - acc: 0.9950 - val_loss: 0.0082 - val_acc: 0.9897\n",
            "Epoch 15/20\n",
            "42000/42000 [==============================] - 2s 50us/step - loss: 0.0040 - acc: 0.9957 - val_loss: 0.0074 - val_acc: 0.9909\n",
            "Epoch 16/20\n",
            "42000/42000 [==============================] - 2s 51us/step - loss: 0.0036 - acc: 0.9963 - val_loss: 0.0074 - val_acc: 0.9907\n",
            "Epoch 17/20\n",
            "42000/42000 [==============================] - 2s 52us/step - loss: 0.0032 - acc: 0.9968 - val_loss: 0.0074 - val_acc: 0.9910\n",
            "Epoch 18/20\n",
            "42000/42000 [==============================] - 2s 50us/step - loss: 0.0029 - acc: 0.9970 - val_loss: 0.0068 - val_acc: 0.9915\n",
            "Epoch 19/20\n",
            "42000/42000 [==============================] - 2s 49us/step - loss: 0.0026 - acc: 0.9976 - val_loss: 0.0070 - val_acc: 0.9914\n",
            "Epoch 20/20\n",
            "42000/42000 [==============================] - 2s 51us/step - loss: 0.0025 - acc: 0.9976 - val_loss: 0.0067 - val_acc: 0.9915\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wQRXxavVyMg0",
        "colab_type": "code",
        "outputId": "a759be30-ec4d-4657-ad44-f8400d437d02",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 587
        }
      },
      "source": [
        "# Training History Visualization\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "# Plot training & validation accuracy values\n",
        "plt.plot(history.history['acc'])\n",
        "plt.plot(history.history['val_acc'])\n",
        "plt.title('Model accuracy')\n",
        "plt.ylabel('Accuracy')\n",
        "plt.xlabel('Epoch')\n",
        "plt.legend(['Train', 'Test'], loc='upper left')\n",
        "plt.show()\n",
        "\n",
        "# Plot training & validation loss values\n",
        "plt.plot(history.history['loss'])\n",
        "plt.plot(history.history['val_loss'])\n",
        "plt.title('Model loss')\n",
        "plt.ylabel('Loss')\n",
        "plt.xlabel('Epoch')\n",
        "plt.legend(['Train', 'Test'], loc='upper right')\n",
        "plt.show()"
      ],
      "execution_count": 49,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZAAAAEdCAYAAAAikTHKAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nGBMRSyDGGGMi8v8B7HYS4qLI5iUAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZAAAAEdCAYAAAAikTHKAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nB5HSgUsxPlaqAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QDhcFRTp0ex2",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "y_predtest_mlp = model.predict(X_test_mlp)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vzESWl0s1c4A",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "y_pred_mlp = model.predict(X_train_mlp)\n",
        "prec_mlp, rec_mlp, thresh_mlp = precision_recall_curve(y_train_5, y_pred_mlp)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Q7ArkWbH2WRO",
        "colab_type": "code",
        "outputId": "91b1d3b9-90b8-4a0d-f9fe-6bd3fd594ad2",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 292
        }
      },
      "source": [
        "plot_precision_vs_recall(prec_mlp,rec_mlp)\n"
      ],
      "execution_count": 52,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZMAAAETCAYAAADzrOu5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9eSYrOBx3KZ8",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "fpr_mlp, tpr_mlp, threshs_mlp = roc_curve(y_train_5, y_pred_mlp)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uokmftv93aMa",
        "colab_type": "code",
        "outputId": "0e497406-02c4-4e22-c05d-a12b24befa79",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 308
        }
      },
      "source": [
        "plot_roc_curve(fpr_mlp,tpr_mlp)\n",
        "plt.axis([-0.008,1,0.0,1.004])\n",
        "\n",
        "roc_auc_score(y_train_5,y_pred_mlp)"
      ],
      "execution_count": 54,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.99770487755197"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 54
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZQAAAESCAYAAADaLCNlAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0\ndHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAAgAElEQVR4nO3deXgUVdbA4d9JwhL2PSiKrDqyCGJc\nENSgI44jCoqjKCIKEcUFFFwQdVgUUMEFUVEQZWTQT1xQFBVcCIuooAgMizAKssgAspMQQpI+3x/V\ngaZN0t1Jd6rTOe/ztOmqulV1ugx9UvfeuldUFWOMMaa44twOwBhjTGywhGKMMSYsLKEYY4wJC0so\nxhhjwsISijHGmLCwhGKMMSYsLKEYY4wJixJPKCJyt4j8ICJZIjI1QNn7RGS7iBwQkddFpEIJhWmM\nMSZEbtyhbAOeAF4vrJCIXAYMAS4BTgGaACMiHp0xxpgiEbeelBeRJ4CTVPWWAra/BfymqkO9y5cA\n01W1fqBj16lTRxs1akRGRgaVK1cOZ9hRR4/+BxTF4zm6eHSDAocOHaJSYiW/LccWFPCoOkdRnzLH\nbfPZ70+/Npr/ar/yhf22KaB67PwFHKJw6lyHghw5kk358uUKP6Z3pSeIfxvF/tej+b4t4qGCOIJP\nkZycHBISEgIV+/O2wi5akMcoXuHw7p6bm0t8fHzxAqCg6xIpxbzeBdi3ed0uVa1b1P3z/22KDi2B\nj3yWVwBJIlJbVXf7FxaRfkA/gKSkJMaNG0d6ejpVqlQJe2Cqzj/dXAWP95Xrgcwcdd4r7DzkYe9h\nJS7O2bbzkFLO+zub64HNBzzUrCjeYyi5Cr/u81A3UVi310NSJTnu+B5VDhyBeMn7gg897gPhvAgR\nJCGuD0ViGI5hTKxQVTLWpJHYuB3xlaqz76kum4pzvGhOKFWA/T7Lee+rAn9KKKo6CZgEkJycrCkp\nKaSlpZGSklLgCbJyctmxP4tNezL4dWc6a/93kK37DpFYLp5Vvx8gqXpF1mzbT6XyCZRPiCPXo2Tn\nejh4OCdsH9LfnsNOpthxKP+Mkeu3Ok4gPk4QEeJFyMzOpXbl8sTHCXEiiDhfxFlZWSQmVkQEZz0g\nPttFhPLxcVQoF0e8CHFxzvHi45z3ieXiSIiLAzn2xS5Hj+Nd9lmXt0IQv+1+68TZkrccJ5BYLj7f\n4xwrj995/7wtb+FYrE65337bSOPGjf+8TY6e7ei6CgnxJMTJceuOK+O/Qz7r/7xP/sfzT5bHb8u/\n4J/3CTIG7/LatT/T4vTTCyznf+w4n/93x66/HP2dyne/fI9V0Dny27+gwsGeK3BcK1eu4Iwz2hQc\nVz5HLrhscCsL+lz5X4MCzpVP4VDi2rFtK6OHDmLx/K+4/b4h3Nb/AZKfKuBkQYrmhJIOVPNZznt/\nsCgHO3QkhwXrd/HM3HX8d2d6UPtsP3AYgP2Z2fluLxfvfOEmxMURHyfsz8wmIU5oWKsSR3I9ZOV4\nODWpCqfUrkz5+Dh2HDhM86SqVCwXR/n4OPYeOkLjOlWOHidehFxVGtRIJLF8PJXKJRAff+yLPT5O\nqFTe+ZKL837pBytQci0r0tJ+JyWludthRIWa+38h5cwGbofhOs+2BC48tci1PKWOx+PhlVde4aGH\nHkJVmTBhAnfeeSdxccVvUo/mhLIaaAPM8C63AXbkV91VmB837eGJ2Wv5afO+Qss1qJHI6SdU46Sa\niZxQvSKN6lSmasUEqieWo3x8HNUSy5HgTR4J8c4Xe0F//RhjTLQaMWIEI0eO5NJLL2XSpEk0atQo\nbMcu8YQiIgne88YD8SJSEchRVf96pDeBqSIyHadn2KPA1FDO9cmvR3jv82+PLtepUoE2J1Wn3Sk1\nuaZdA+pWqUBCvD2KY4yJbTk5Oezdu5e6detyxx130LhxY3r37h32P4rduEN5FBjms3wTMEJEXgfW\nAC1UdbOqfi4iTwPzcNpS3/fbr1DZuR4+/82pqmrVoBoTbmhH4zqx3ePLGGP8rVy5kr59+5KYmEha\nWhonnHACt9xyS0TOVeIJRVWHA8ML2HxclyxVfRZ4tijnWb3tAOnZ0LBWJT6554KiHMIYY0qtI0eO\nMGrUKEaPHk3NmjV58cUXI15NH81tKMXy8/+cTrKnn1DV5UiMMaZk/frrr3Tt2pXVq1fTs2dPnn/+\neerUqRPx88ZsQtlXQM8sY4yJdfXr16d27dp88sknXHHFFSV23phtka6Y4Hy0ctbobowpA+bPn0+X\nLl04fPgwlStXZv78+SWaTCCGE0rek+R1qth4ksaY2HXgwAH69+9PSkoKa9euZfPmza7FEsMJxcko\n9qiIMSZWffbZZ7Rq1YpXX32V++67j5UrV3Lqqae6Fk/MtqHkKWhICGOMKc08Hg8jRoygatWqLF68\nmPPOO8/tkGI3oeTdoYQwOokxxkS9Dz74gAsuuIC6devy/vvvU6dOHSpUiI6q/Riu8nJ+2vAoxphY\nsH37drp370737t0ZP348AA0aNIiaZAIxfIeiRxOKu3EYY0xxqCpvvvkm9913H4cOHeKpp55i0KBB\nboeVrxi+Q8mr8rKMYowpvcaMGcMtt9xCy5YtWblyJQ8++GCBE6O5LTqjCoO8mSgtnRhjShuPx8O+\nffuoVasWffr0oWbNmtx+++1hGWI+kqI7umLIq/KyOxRjTGmyfv16UlJS6Nq1Kx6Ph/r169O/f/+o\nTyYQwwnFczShuBuHMcYEIycnh6effpo2bdrwn//8hz59+pS6TkUxW+XlsVZ5Y0wpsWnTJq699lp+\n+OEHunXrxssvv8wJJ5zgdlghi9mEkjf1ut2hGGOiXZ06dUhISGDGjBlce+21pe7OJE/MVnmp9fIy\nxkSxJUuWcPXVV5OZmUnlypVZvHgx//jHP0ptMoEYTige6+VljIlChw4d4v7776d9+/YsXbqUDRs2\nALHxEHbMJpTNezIBiLM6L2NMlJg/fz5t2rThmWeeITU1ldWrV9OyZUu3wwqbmG1DqVfVGY5g695M\nlyMxxhinGv7hhx/G4/Hw9ddf06lTJ7dDCruYTSgJ3juThrUquRyJMaYs++yzz0hOTqZu3bq88847\n1KpVi8qVK7sdVkTEbJVXnhioljTGlEK7d++mV69e/P3vf+fpp58G4OSTT47ZZAIh3KGIyHnA34Dz\ngBOBRGAXsA6YD3yoqnsjEaQxxpQWqsq7777L3Xffzd69e3nsscd45JFH3A6rRAS8QxGR3iLyH2Ax\ncB9QCfgv8D2wFzgXeA34XUSmikjjCMZrjDFR7dlnn+X666+nYcOG/PDDD4wcOTKqhpiPpELvUERk\nJVAXeBO4GViueQ94HF+uOtAF6AmsEZFbVPWdCMRrjDFRR1XZv38/NWrUoGfPngAMHDgwakcFjpRA\nn3YK8KqqHi6skKruB6YD00WkDVA/TPEZY0xU27RpE/369ePQoUPMnz+f+vXrM3jwYLfDckWhVV6q\nOj5QMslnnxWqOqd4YRXfn26jjDEmjDweDy+++CItW7bkm2++oUePHm6H5LqYvx+zTl7GmHD7/fff\n6dGjB4sWLaJz585MmjSJU045xe2wXBeoDeWfIRxLVfXxYsZjjDFRr0aNGmRmZvLGG2/Qu3fvmBg2\nJRwC3aEMD+FYClhCMcbEpJUrVzJq1CimTp1K5cqVWbp0qSUSP4HaUOJCeMWXVNDGGFNSsrKyGDZs\nGGeddRZpaWmsW7cOiI3BHMMt5p+UN8aYovr+++8566yzGDlyJDfccANr1qyhbdu2bocVtWK2UT6f\nx2WMMSZoqsrgwYPZv38/s2fP5u9//7vbIUW9kBKKiPQD+gOnAX969DMaq73srtQYE4q0tDRatGhB\nvXr1eOutt6hRowbVqlVzO6xSIegqLxG5GZgALAUqAm8A/wYOAL8CIyMRoDHGlIQDBw5wxx130KlT\nJ0aPHg1Aw4YNLZmEIJQ2lHuBMTh3KAAvq2pvoAmQCewO5iAiUktEZopIhohsEpEbCyhXQUReEZEd\nIrJHRD4WkQYhxGuMMUH59NNPadmyJZMnT2bw4MFHE4oJTSgJpTmwAPB4X+UBvCMMjwIGBnmcl4Aj\nQBLO2F8TRSS/KcsGAu2BM3BGN96Lc4dkjDFhM3PmTK644gqqV6/Ot99+y7hx46hUyeZRKopQ2lAy\ngThVVRHZjnNn8p13WzrOl36hRKQy0B1oparpwCIRmQX0Aob4FW8MzFHVHd593wGeDSFeY4zJl6qS\nnp5O1apVueiii6hbty4PPPBAmRkVOFJCSSj/AZoBXwILgaEishHIwXkA8ucgjnEqkKOq633WrQAu\nyqfsFGC8iJwI7MO5m/msoAN7Owz0A0hKSmLLli0A/PrrBtI8W4IILbalp6eTlpbmdhius+twTFm9\nFrt37+b5559n3759jB8/nvLly9OxY0e+/fZbt0Mr9UJJKJNw7koAHsNJLIu8yweBbkEcowpOI76v\n/UDVfMr+F9gC/A7k4iS0uws6sKpO8sZIcnKynnTSyfDbRpo2bULKhU2DCC22paWlkZKS4nYYrrPr\ncExZuxaqytSpUxk0aBCZmZmMHDmSiy66iIULF5ap6xBJQScU3/lNVPUXb7tHe5wJtxar6q4gDpMO\n+HeZqIaTkPy9hNM1uTaQATyIc4dybrAxA4gND2lMmbdjxw5uvvlm5s6dS8eOHZkyZQqnnnqq22HF\nnCI/Ka+qGar6parOCjKZAKwHEkSkuc+6NsDqfMq2Baaq6h5VzcJpkD9HROoUNWZjTNlUtWpVdu7c\nyYsvvsj8+fMtmURIyE/Ki0h9oCHOsyjHUdUFhe2rqhki8gEwUkRScZJGV+D8fIovBW4WkTTgEHAn\nsC2E5GWMKcPWr1/P448/zquvvkqlSpX48ccfiYuz0aYiKZQHGxuIyNc4bRrfAvN8Xmnen8G4E0gE\ndgJvA/1VdbWIXCAi6T7l7gcO47Sl/AH8Hbg62HiNMWVTTk4OTz31FGeccQazZ89m9WqnAsSSSeSF\ncocyEWiN05bxHyCrKCdU1T3k04CvqgtxGu3zlnfj9OwyxpigrFixgr59+/Ljjz9yzTXX8NJLL1G/\nvs1IXlJCSSgXAANUdVqkgjHGmOIYOHAgW7Zs4d133+Xaa691O5wyJ9QHG3dGKpBws7GGjSkbvv/+\nexo3bky9evWYOnUqVatWpXbt2m6HVSaFUqk4GeeJ9lLFRhs2JjZlZGQwaNAg2rdvz4gRIwBo1KiR\nJRMXhXKH8jvQS0S+wnkeZI9/AVV9PVyBGWNMQebNm0dqaiobNmzgjjvuYMyYMW6HZAgtobzi/dkI\n6JTPdgUsoRhjImrKlCmkpqbStGlT5s2bZ0+5R5FQEkrjiEVhjDEBZGZmkpiYSJcuXRg6dCiPPPKI\njQocZUIZemVTJAMxxpj87Nq1i3vvvZeNGzeyYMECkpKSGDVqlNthmXzYkz7GmKikqsyYMYMWLVow\nY8YMLr30Ujwej9thmUIUeociIhuAq1V1hXeo+sJ646qqRs2wvmr9ho0ptXbv3k1qaioffvghZ599\nNlOmTKF169Zuh2UCCFTlNZ9jw83Pxx7vMMaUgMTERH799VfGjh3LvffeS0JCyMMOGhcU+n9JVW/1\neX9LxKMxxpRZv/32G48//jgTJkygUqVKLFu2zBJJKVPsNhQRsaeIjDFF5vF4mDBhAq1atWLGjBn8\n9NNPAJZMSqFQRhu+TUQe8FluLSJbgZ0i8oN3WHtjjAnaunXruPDCCxkwYAAdO3Zk1apVdOjQwe2w\nTBGFcodyD854XnmexZnr/V6gOjAyjHEZY8qAu+66izVr1vCvf/2Lzz77jFNOOcXtkEwxhHJPeQrw\nM4CIVAcuArqp6qcishuwsQ+MMQGtWLGC+vXrk5SUxOTJk0lMTLQh5mNEKHcocUBeJ/COOD2+0rzL\nW4B64Qur+NQ6pBkTVbKysnjsscdITk7m0UcfBaBx48aWTGJIKHco/wWuAL4GegCLVfWQd9uJ5DNY\nZDQQG27YGNd999139O3blzVr1nDzzTfz5JNPuh2SiYBQEso4YJqI9AZqAv/w2dYJWBnOwIwxsWH6\n9On06tWLBg0a8Omnn3L55Ze7HZKJkFDG8npLRDYD5wJLVXWBz+YdwKxwB2eMKb2ysrKoUKECl156\nKffddx/Dhg2jWrVqbodlIiikjt6qughYlM/6YWGLyBhTqu3fv58HH3yQVatWsWDBAurVq8czzzzj\ndlimBNjgkMaYsPnkk09o2bIlr732Gueffz45OTluh2RKkCUUY0yx7du3j5tuuokrr7ySmjVr8t13\n3zF27FgqVKjgdmimBMVsQrHRho0pOeXLl+enn35i+PDh/Pjjj5x99tluh2RcEPOD5VinYWMiY9u2\nbTzxxBOMGzeOSpUqsXz5csqVK+d2WMZFMXuHYoyJDFVlypQptGjRgjfeeIMlS5YAWDIxllCMMcHb\nuHEjnTt3JjU1lTZt2rBy5UpSUlLcDstEiZASioj0E5GfROSQiOT6vyIVpDEmOvTv35/vvvuOl19+\nmXnz5tG8eXO3QzJRJOg2FBG5GZgA/AtoA7wOlAOuAv4ApkciQGOMu9atW0eNGjVISkri5ZdfJiEh\ngYYNG7odlolCodyh3IszonB/7/LLqtobaIIzrP3uMMdmjHFRdnY2Y8aMoU2bNgwZMgSAJk2aWDIx\nBQoloTQHFuCMOOwBygOo6l5gFDAw7NEZY1yxfPlyzj33XIYOHUqXLl0YM8ZmpzCBhZJQMoE4VVVg\nO86dSZ50nBGHo44NNmxMaN59913OPvtstm3bxnvvvcd7771nQ8yboISSUP4DNPO+XwgMFZH2InI2\nMBzv5FvGmNIpOzsbgJSUFPr168eaNWvo3r27y1GZ0iSUhDIJZ9h6gMeAKjgDRX4HnAoMDm9oxpiS\nkJGRwb333ktKSgq5ubnUrVuXl156iVq1arkdmillgk4oqvqOqo7xvv8FaAlcBlwNNFPVtGCPJSK1\nRGSmiGSIyCYRubGQsu1EZIGIpIvIDhGxthpjwuSrr76idevWjB8/nrZt23LkyBG3QzKlWJGHXlHV\nDODLIu7+EnAESALaArNFZIWqrvYtJCJ1gM+B+4D3cDoCnFTUmI0xjoMHDzJu3Dhmz55N8+bNmT9/\nPhdeeKHbYZlSrtCEIiINgf+parb3faFUdXOgMiJSGegOtFLVdGCRiMwCegFD/IoPAuaoat4zLlnA\n2kDn8MYSTDFjyqSEhARWrVrFgw8+yPDhw0lMTHQ7JBMDpLAvXu/T7+1VdYmIeIBCv6VVNT7gCUXO\nBL5R1Uo+6+4HLlLVK/3Kfo3TGeBsnA4B3wN35Ze4RKQf0A8gKSnprCv++S++2pxDz9PLc+kpNsZQ\neno6VapUcTsM15Xl67Bv3z6mT59Onz59SExMZM+ePdZOQtn+nfDXqVOnH1U1uaj7B6ry6gP86vM+\nHH/2VwEO+K3bD1TNp+xJQDvgUpzE8jTwNtDBv6CqTsLpOEBycrKe1KABbN5E82bNSOnQOAxhl25p\naWk25hJl8zqoKu+88w733HMP+/fvJzU1lZSUlDJ5LfJj1yF8Ck0oqvovn/dTw3TOdMB/YulqwMF8\nymYCM1V1KYCIjAB2iUh1Vd0fpniMiVnbtm2jf//+zJo1i7PPPpvXX3+dVq1auR2WiVFB9/ISkboi\ncmoB2071NqAHYz2QICK+o8q1AVbnU3Ylx98VWcOIMSHo168fc+fOZezYsSxevNiSiYmoUHp5vQzs\nAW7PZ9t9QG3gukAHUdUMEfkAGCkiqTi9vLoC5+dT/A3gfRF5ASfhPAYssrsTYwq2ceNGKlWqRFJS\nEuPHj8fj8diowKZEhPJgY0dgTgHb5pJPu0Yh7gQSgZ04bSL9VXW1iFwgIul5hVT1a2AoMNtbthlQ\n4DMrxpRlHo+HF154gVatWnH//fcD0LRpU0smpsSEcodSE6fxPD8HcO5QgqKqe4Bu+axfiNNo77tu\nIjAx+DC9+4W6gzGl2M8//0zfvn1ZvHgxl19+OaNHj3Y7JFMGhXKHshU4t4Bt5wL/K3444Sc2OqSJ\ncbNmzaJt27b8/PPPvPnmm8yePZuTTz7Z7bBMGRRKQnkPeFhErvBd6V0eAswIZ2DGmMLl5jqTpLZv\n356bbrqJNWvW0KtXL/sjyrgmlIQyEudZkFki8ruILBGR34FZ3vUjIhGgMeZ4hw8f5pFHHjluMMfX\nXnuNpKQkt0MzZVwog0MeAi4CbsOZaGsfMB/oi/OU+6GIRGiMOWrx4sWceeaZjB49mmbNmnH48GG3\nQzLmqJAGh1TVbJy55F+PTDjGmPwcOnSIoUOH8sILL3DyySfz+eefc9lll7kdljHHCaXKyxjjEhHh\n888/584772TVqlWWTExUCuVJ+fIiMkxEfhaRQyKS6/fKiWSgobLBhk1pt3//foYMGUJGRgaJiYks\nW7aMF198kapV8xv2zhj3hVLlNRa4C/gM+ABnKPmoZx1eTGn08ccfc8cdd7B9+3Y6dOjAlVdeSaVK\nlQLvaIyLQkko1wLDVHVUpIIxpqz7448/GDhwIG+//TatW7fmo48+Ijm5yKOJG1OiQmlDqQJ8G6lA\njDHOYI7vvfceI0aM4IcffrBkYkqVUBLKx4DNEWpMmG3bto0dO3YAMHbsWJYtW8Y///lPypcv73Jk\nxoQmlIQyAbhBRP4pIski0sT/FakgjYlFqsqUKVNo0aIF9957LwDNmjWzIeZNqRVKG0pedddwYFgB\nZQJOAWyMcYaY79evH19++SUpKSk8/vjjbodkTLGFklDCNQVwidDSE6opY+bMmcM111xDfHw8r7zy\nCrfddhtxcfZImCn9gk4oYZwCuERZr2ETLTweD3FxcbRr146rr76aMWPG2KjAJqaE/GeRiMSJSCsR\nuUhEKkciKGNiSXZ2NqNHjz5uMMd///vflkxMzAkpoYjIXcB2nLnevwZO867/UEQGhD88Y0q3n376\niXPOOYdHHnmEpKQkMjIy3A7JmIgpNKF470bu976/DRgPfIgzd7xvbdJCoHukgjSmtMnKyuKRRx7h\n7LPPZvv27bz//vu8++67VKtWze3QjImYAhOKiLQDfgDaeVcNAp5R1X7ATL/iP+O9WzHGOF2C33//\nfXr16sWaNWu45ppr3A7JmIgr7A7le2Caqt7oXW4MzCmgbAZQI5yBGVPaZGRkMHz4cNLT06lYsSJL\nly7ljTfeoGbNmm6HZkyJKCyhvA7cKyIXeZd3AY0KKHsa8HsY4yo2G23YlKQvv/ySVq1aMWLECObM\ncf7uslGBTVlTYEJR1duBm4AXvas+Af7p90S8ikgd4D6ctpXoY8MNmwjat28fqampXHrppZQvX56F\nCxfSvbs1J5qyqdBGeVVdCJzlXXwUZ8j6VcCXOA85vgCsBXJx5pw3pky5/fbbmTp1Kg899BDLly+n\nY8eObodkjGsCdhtW1SPen7uAZGAMUA74FefByBeB9qq6P4JxGhM1/vjjj6ODOY4ePZrvv/+eJ598\nksTERJcjM8Zdoc4pfxB43PsypkxRVd5++20GDBjAxRdfzIwZM2jatKnbYRkTNWwAIWOC8Pvvv3PV\nVVfRs2dPmjZtyrBhBY2PakzZVegdioh8Ddypqj973xdGVfWS8IVmTHSYN28e3bp1Izs7m2eeeYaB\nAwcSH28DaxvjL1CVl28XqTgKH204qrpTWa9hU1yqiojQunVrOnfuzJgxY2jWrJnbYRkTtQpNKKra\nyed9SsSjiYCoynKmVMjNzWXChAnMmjWLL774gjp16vDuu++6HZYxUc/aUIzxsXbtWi644ALuu+8+\nEhMTOXjwoNshGVNqBJ1QROQhEZlQwLYXROSB8IVlTMnKzs5m1KhRtG3blnXr1jFt2jQ++eQTatSw\nEYWMCVYodyi34gxbn5/l3u3GlEq5ublMmzaNrl27smbNGm666SbERlkwJiShJJSGwH8L2LYBOKX4\n4RhTcg4fPsyTTz55dDDHb7/9lhkzZpCUlOR2aMaUSqEklENAgwK2nYQzLEtQRKSWiMwUkQwR2SQi\nNwYoX15E1orI1mDPYYNDmsIsXryYM888k4cffphZs2YB2KjAxhRTKAllIfCAiFTwXeldHuzdHqyX\ngCNAEtATmCgiLQsp/wDwRwjH94mvKHuZWJWZmcnAgQPp2LEjmZmZzJkzhxtvLPTvGWNMkEIZemU4\nsBhYLyL/xhmuvgHOiMS1gVuCOYh3HvruQCtVTQcWicgsoBcwJJ/yjb3nGARMDiFeY/7kueee48sv\nv+Tuu+9m9OjRVKlSxe2QjIkZoiHUDYnIOcA44HycuxsPsAi4X1V/CPIYZwLfqGoln3X3Axep6pX5\nlP8EmALsBf6tqicVcNx+QD+ApKSks/726L9I25pD7xbl6dSwXNCfMValp6eX2S/P9PR0srOzqVmz\nJuvXrycrK4vWrVu7HZbryvLvhC+7Dsd06tTpR1VNLur+oQ4OuQS4UEQSgZrAXlXNDPGcVYADfuv2\nA3+ajUhErgbiVXWmiKQEiG0SMAkgOTlZTzjxRNi6mVNPO5WUc62/QFpaGikpKW6HUeI++ugj+vfv\nT/v27Xn//fcByuR1yE9Z/Z3wZ9chfIr0YKOqZqrqtiIkE4B0oJrfumrAcU+QeavGngYGFCVGU7b9\n8ccf9OjRg27dulG3bl2GDh3qdkjGxLxAg0P+E3hNVbd53xdGVTWYYe3XAwki0lxV87ohtwFW+5Vr\njjPl8ELv8wDlgeoish04T1V/C+Jcpgz65ptv6Nq1KwcOHGDkyJE89NBDlC9f3u2wjIl5gaq8hgOf\nA9u87wujBDFPiqpmiMgHwEgRSQXaAl1x2mV8rQJO9lk+H2cyr3YE1ePL+g2XNXmDOf7lL3+hffv2\nPPnkk7RsWVjnQWNMOAWaAjjO226S976wVyjjed8JJAI7gbeB/qq6WkQuEJF07/lyVHV73gvYA3i8\ny7nBnkhseMiYp6pMnjyZzp07k5OTQ+3atfn4448tmRhTwgpNKCKyLO/5EBH5p4icGI6TquoeVe2m\nqpVVtaGqvuVdv1BV8+1uoappBfXwMmXXhg0b+Otf/0q/fv3Iyclh3759bodkTJkVqFG+NU6vLIBh\nOE/EG+O63Nxcnn/+eVq3bg72wv0AABexSURBVM3SpUt59dVX+eqrr6hTp47boRlTZgVqQ9kGdBOR\nHThTi9QXkYYFFVbVzeEMzpiCZGdn8+qrr5KSksIrr7zCySefHHgnY0xEBUoorwJPAA/itHLPDFDe\n5kU1EZOdnc2LL75IamoqVatWZeHChdSuXdtGBTYmSgSasXG0iHwBtADeAMbgjCxsTIlatmwZffr0\nYcWKFdSoUYNbb73VqreMiTIBn5RX1aXAUhG5BZimqj9HPKowsNGGY8Phw4cZMWIEY8eOpV69esyc\nOZNu3bq5HZYxJh9BD73iO798aWK1IaXbXXfdxeuvv06fPn0YN26cDTFvTBQLaegVETlTRD4QkV0i\nkiMi7bzrR4vI3yIToilr0tPT+eMP59nVoUOHMmfOHKZMmWLJxJgoF8qc8h2Bb4G/AG/57esB7ghv\naKYs+uKLL2jdujX9+vUDoGnTpnTu3NnlqIwxwQjlDuVJYA7QEmduEl/LcIZEMaZI9u3bR9++fenc\nuTMVKlRg8ODBbodkjAlRKMPXtwOuUVUVEf8m711A3fCFZcqSJUuW0K1bN3bu3MmQIUMYNmwYFStW\ndDssY0yIQkkoh4FKBWw7AWdOE2NC1rRpU1q1asWYMWM466yz3A7HGFNEoVR5LQLuFRHfhxfz7lT6\nAl+HLaowsG7D0UtVeeutt7j88suPDuY4d+5cSybGlHKhJJTHcKq9VnjfK9BbROYB5wEjwh9e8Vmv\n4eiydetWrrrqKnr27Mm+ffvYvXu32yEZY8Ik6ISiqiuAC4EdwCM439V3ezdfpKrrwh+eiRV5Q8y3\nbNmSr7/+mueee45FixaRlJTkdmjGmDAJdU75ZcAlIlIRqAXsU9VDEYnMxJQjR47w3HPPkZyczOTJ\nk2nSpInbIRljwiykhJJHVQ/jjERsTIFyc3OZNGkSN910E1WrVmXevHnUq1fPBnM0JkYFmmBrkPdu\nJGgi0s6emjerV6+mQ4cO3Hnnnfz73/8GICkpyZKJMTEsUBtKL2CjiDwpIm0KKiQiNUWkl4jMxekN\nVi2cQZrSIzs7m8cff5x27drxyy+/MH36dO64wwZRMKYsCFTl1Q4nqQwGHhSRA8B/gD+ALKAm0ARo\n6l1+B2ihqr9FKuBgKdZv2A0DBw5k4sSJXH/99bzwwgvUq1fP7ZCMMSUk0HwoCrwJvCki5wJ/A87F\nSSIVgd3AQmAU8JGqRt2E3lbDEnmZmZkcOnSI2rVrM3jwYDp37mxDzBtTBoUyfP33wPcRjMWUQosW\nLaJv376cdtppzJo1i6ZNm9K0aVO3wzLGuCCk4euNyZOens4999zDhRdeSFZWFvfcc4/bIRljXFak\nbsOmbFu+fDndunVj8+bN3H333YwePZoqVaq4HZYxxmWWUEzIGjZsSKNGjZg+fTodOnRwOxxjTJSw\nKi8TlA8//JArr7ySnJwcatWqRVpamiUTY8xxYjah2GjD4bFz506uv/56rr76arZu3cqOHTvcDskY\nE6XCllBEpEK4jhVOYuMNF4mqMn36dFq0aMGHH37IE088wZIlS2jQoIHboRljolSx21C8Q7P0B+4H\n7NsmRmRnZzN69GiaN2/OlClTaNGihdshGWOiXMCEIiINgR5AQ+AX4A1V3S8i5YF7gAeAesC3kQzU\nRJ7H4+HNN9+ke/fuVK1alblz51K/fn3i4+MD72yMKfMKTSgi0gGYzfFjc90uIlcBHwAtgWVAH1X9\nNGJRmoj79ddfSU1NJS0tjf379zNw4ECr3jLGhCRQG8ownOFVOuLMJ98SZ4KtxTjDr9yqqsmWTEqv\n3Nxcnn32WVq3bs2yZcuYPHkyAwYMcDssY0wpFKjK62xggKou9i6vFZE7gVXAQFX9V0SjMxE3ePBg\nxo8fT5cuXZg4cSInnXSS2yEZY0qpQAmlOvBfv3V5y0vCH074WK/hgmVnZ3Pw4EFq1arFgAEDOOec\nc7jhhhtsrhJjTLEE020412/Z4/15pKgnFZFaIjJTRDJEZJOI3FhAuQdEZJWIHBSRjSLyQOgnK2qU\nsenHH38kOTmZ3r17o6o0adKEG2+80ZKJMabYguk2PEJEdvks533zPC4ie3zWq6r2DvK8L+EkpCSg\nLTBbRFao6mq/cgLcDKzEmXNlrohsUdX/C/I8xisrK4shQ4Ywbtw4kpKSePzxxy2JGGPCKlBC2Qyc\nns/6TTgN9L6CqmUSkcpAd6CVqqYDi0RkFs5EXkOOO6Dq0z6L60TkI6ADYAklBGvWrOG2225jy5Yt\npKamMnbsWGrUqOF2WMaYGBNogq1GETjnqUCOqq73WbcCuKiwncT5c/oC4NUCtvcD+oEzd/n27dsB\nWPfzz6Sl/xqGsEuvAwcOULlyZcaNG8dZZ53F8uXL3Q7JNenp6aSlpbkdRlSwa+Gw6xA+bow2XAU4\n4LduP1A1wH7Dcdp83shvo6pOAiYBJCcna/369eH3rZz2l7+Qknxy8SIuhebOncvEiROZMWMG5cqV\no2rVqnTq1MntsFyXlpZGSkqK22FEBbsWDrsO4ROwUV5EbhGR5SKSLiJbReQZ71PyRZXO8Q9K4l0+\nWEgMd+O0pVyhqlnBnKSsDg65d+9ebr31Vi677DLWrl3Ltm3bAKy9xBgTcYUmFG/vq9dxHmqcDWwD\n7gVGF+Oc64EEEWnus64N4N8gnxdDH5y2lUtUdWuoJytLX6MzZ86kRYsWTJs2jYcffpjly5dzyimn\nuB2WMaaMCFTlNRCYCVynqrkAIjIMGCIiD+WtC4WqZojIB8BIEUnF6eXVFTjfv6yI9MRJXp1UdUOo\n5ypLcnJyeOyxx6hfvz6zZ8+mXbt2bodkjCljAlV5nQpM9kscLwMVcAaLLKo7gURgJ/A20F9VV4vI\nBSKS7lPuCaA2sNRb5ZYuIq8U47wxRVV55513OHDgAAkJCXz66acsWbLEkokxxhWBEkp1YI/furzl\nmkU9qaruUdVuqlpZVRuq6lve9QtVtYpPucaqWk5Vq/i87ijqeWPJli1b6NKlCz169OCVV5wc27Bh\nQ8qVK+dyZMaYsiqYXl5xIuKbeOILWI+qejAR5fF4mDx5Mg888AC5ubk899xz3HPPPW6HZYwxQSWU\nbwpY/73fsgZ5PFMMQ4cO5amnnuLiiy9m8uTJNGnSxO2QjDEGCJwARlJKx1nU0hl2vnJzczl48CA1\natTg9ttvp1mzZvTt29e6AhtjokqgJ+WHl1AcEVPav3RXr15Nnz59qFOnDp988gmNGzcmNTXV7bCM\nMeZPAj2HskFE2pRUMOaYI0eOMHLkSM4880w2bNhAr1693A7JGGMKFajKqxFOF2FTgtavX88//vEP\nVq5cyQ033MD48eOpW7eu22EZY0yhrBE9CtWtW5dy5coxa9YsrrzySrfDMcaYoAQzwVbstG5HsYUL\nF3LdddeRnZ1NzZo1Wbp0qSUTY0ypUpQJtgoSygRbxuvgwYMMGTKEl19+mcaNG7N582aaNm1a6jsT\nGGPKnmASSlsgmBF+o+tOJrqiydecOXPo168fW7ZsYeDAgYwaNYrKlSu7HZYxxhRJMAmlm6ouiXgk\nERKtf+fn5uZy//33U6lSJRYtWsT55/9pbExjjClVrFG+hM2aNYuUlBSqVavGxx9/TP369alYsaLb\nYRljTLEF0yhvwmDHjh1cd911dO3alRdeeAGARo0aWTIxxsQMSygRpqpMmzaNFi1a8NFHHzFq1Cge\neught8MyxpiwCzT0iiWcYhoxYgQjRoygffv2TJkyhdNPP93tkIwxJiKsDSUCPB4P6enpVKtWjd69\ne1OrVi3uuusu4uPjA+9sjDGlVMzegbjVa/iXX37h4osvpkePHqgqjRs3ZsCAAZZMjDExL2YTSp6S\nej4wNzeXZ555hjPOOIPly5fTvXv3kjmxMcZECavyCoONGzfSo0cPlixZwlVXXcXLL79MgwYN3A7L\nGGNKlCWUMKhRowaHDx/m7bff5vrrr7dhU4wxZVLMV3lFytKlS+nZs+fRwRyXL19Ojx49LJkYY8os\nSyghyszM5MEHH+S8884jLS2NDRs2AKV/ZkhjjCkuSyghWLBgAW3atGHs2LH06dOH1atXc9ppp7kd\nljHGRIWYbUNRDW/HYY/Hwz333ENOTg5ffvkll1xySViPb4wxpV3MJpQ8xa2J+uKLLzjnnHOoXr06\nH3zwAfXr17ch5o0xJh9W5VWAPXv2cMstt9C5c2eeffZZAJo2bWrJxBhjChDzdyhF8cEHH3DnnXey\ne/duHn30UYYOHep2SMYYE/Usofh58sknefjhhznzzDP5/PPPadu2rdshGWNMqWAJBacBPyMjgypV\nqhx9lmTQoEGUK1fO7dCMMabUKPNtKJs3b+aKK67guuuuQ1Vp1KgRDz30kCUTY4wJUcwmlECdhj0e\nDxMnTqRly5YsWLCAyy+/POxdjY0xpiyJ+Sov4c/9hrdu3UrPnj1ZsGABf/3rX5k8eTKNGjUq+eCM\nMSaGxHxCyU+VKlXYtWsXU6ZM4dZbb7VhU4wxJgxitsrL36pVq7j11lvJzs6mRo0arFy5kj59+lgy\nMcaYMHEloYhILRGZKSIZIrJJRG4soJyIyFMistv7ekpCzADZ2UcYMWIE7dq1Y/bs2axfvx7AZlA0\nxpgwc6vK6yXgCJAEtAVmi8gKVV3tV64f0A1og9PO/gWwEXglmJNk/W89D/S8n02//MyNN97I+PHj\nqVOnTtg+hDHGmGNKPKGISGWgO9BKVdOBRSIyC+gFDPEr3ht4RlW3evd9BriNIBKKqrJnzktU1Qw+\n/vhjunTpEtbPYYwx5nhS0l1lReRM4BtVreSz7n7gIlW90q/sfqCzqn7vXU4G5qlq1XyO2w/njgbg\nNGAdUAfYFZEPUvrYtXDYdTjGroXDrsMxp+X3/RosN6q8qgAH/NbtB/L7EFW823zLVRERUb9MqKqT\ngEm+60TkB1VNLn7IpZ9dC4ddh2PsWjjsOhwjIj8UZ383GuXTgWp+66oBB4MoWw1I908mxhhj3OdG\nQlkPJIhIc591bQD/Bnm869oEUc4YY4zLSjyhqGoG8AEwUkQqi0gHoCswLZ/ibwKDRKSBiJwIDAam\nhnC6SYGLlBl2LRx2HY6xa+Gw63BMsa5FiTfKg/McCvA6cCmwGxiiqm+JyAXAZ6paxVtOgKeAVO+u\nrwEPWZWXMcZEH1cSijHGmNhTZoZeMcYYE1mWUIwxxoRFqU8oJTkuWDQL4To8ICKrROSgiGwUkQdK\nOtZIC/Za+JQvLyJrRWRrScVYEkK5DiLSTkQWiEi6iOwQkYElGWskhfBvo4KIvOL9/HtE5GMRaVDS\n8UaKiNwtIj+ISJaITA1Q9j4R2S4iB0TkdRGpEMw5Sn1C4fhxwXoCE0WkZT7lfMcFOwO4Eri9pIIs\nAcFeBwFuBmoCfwPuFpEeJRZlyQj2WuR5APijJAIrYUFdBxGpA3wOvArUBpoBc0swzkgL9vdhINAe\n5/vhRGAvMKGkgiwB24AncDpEFUhELsMZBusS4BSgCTAiqDOoaql9AZVxflFO9Vk3DXgyn7KLgX4+\ny32B79z+DCV9HfLZ9wVggtufwa1rATQG1gKXA1vdjt+N6wCMBqa5HXMUXIeJwNM+y1cA69z+DBG4\nJk8AUwvZ/hYw2mf5EmB7MMcu7XcopwI5qrreZ90KIL+/Plp6twUqVxqFch2O8lb5XUBsPSwa6rWY\nAAwFMiMdWAkL5TqcB+wRkcUistNb1dOwRKKMvFCuwxSgg4icKCKVcO5mPiuBGKNNft+VSSJSO9CO\npT2hhGVcsAjFVpJCuQ6+huP8DrwRgZjcEvS1EJGrgXhVnVkSgZWwUH4nTsIZ2Xsg0BBnioi3Ixpd\nyQnlOvwX2AL87t3ndGBkRKOLTvl9V0Lg75NSn1BsXDBHKNcBcBrocNpSrlDVrAjGVtKCuhbeaRSe\nBgaUUFwlLZTfiUxgpqouVdXDOPXl54tI9QjHWBJCuQ4vARVw2pEq44zoURbvUPL7roRCvk/ylPaE\nYuOCOUK5DohIH7yNbuqdayaGBHstmgONgIUish3ny+MEb8+WRiUQZ6SF8juxEmcCuzyx8EdWnlCu\nQ1uctoU93j+yJgDneDstlCX5fVfuUNXdAfd0u4EoDA1M/4dze14Z6IBze9Yyn3J34DS+NsDpwbEa\nuMPt+F24Dj2B7cDpbsfs5rXAmbqhvs/rGpxeMPVxqsFc/xwl+DtxMU6PprZAOeA5YKHb8btwHd4A\n3geqe6/DUOB3t+MP43VIACoCY3A6JlQEEvIp9zfvd0QLoAbwNUF08FHVmEgotYAPgQxgM3Cjd/0F\nOFVaeeUEp4pjj/f1NN6hZ2LhFcJ12Ahk49zW5r1ecTt+N66F3z4pxFAvr1CvA9Afp+1gL/AxcLLb\n8Zf0dcCp6poO7AT2AYuAc9yOP4zXYTjO3afvazhOu1k60NCn7CBgB05b0htAhWDOYWN5GWOMCYvS\n3oZijDEmSlhCMcYYExaWUIwxxoSFJRRjjDFhYQnFGGNMWFhCMcYYExaWUEypJiK3iIgW8Pqrt0wj\nv/VHRGS9iDwnIjV9jjXcr1yWiKzxziET9L8VERkkIit9x4nzrvtYRP7nPfbwsF6IY+dpLyLfe+f+\nUBFpm8/nyvHOCzLFf74PEXleRD6NRGwm9iW4HYAxYfIPwH8YmTV+y2OAWTjjNXUAHgXOFJFOevwD\nWR2BXJwH4m7BeQjWAzwTKAgRqQE8Atzud8zbcB4S+xBn1IZImYIzNteVwCGcoUfy5H2ucjhPQY8A\nzhKRdqrq8ZZ5CtjgvSbzIhiniUGWUEysWK6qvwQos0FVv/O+ny8i5XCeFD4TWOZT7ntVzQEQkc9x\nJly6jSASCs48O0cA/xGMW6qqR0QSiFBC8d5FnQaMUtWvfdbnvT36uXDGMMsFJnv3WQugqv8TkY9x\nJh2zhGJCYlVepixb6v3ZrKAC3r/cV+AMTxGMVGCGqubmc5wiE5FK4kxbvdFbZbdRRB7Jq4oTkVtw\n7j7igMe8VVu/BThs3rDu5fzW/x9wmYicXJyYTdljCcXEingRSfB5xQexT2Pvz30ByjUCfg10MBE5\nBfgLsDCIcwfNe1czBydZjceZXfI14DFgrLfYbJwqLXCqvdoDV/sdKu8aJYrIWTiDH64GVvmVW4jz\n3XBpOD+HiX1W5WVixc9+y99w7As2T5z3y7k8x9pQ/sefE0C8t5qoJs6X+FnAtUHEcJ7354pCS4Xu\nBpzPcpGqLvCu+8ob4zAReUpVd4rIXu+2rT5Ve74O+y3/DHTxv3tS1T9EZCvO5yl0/nFjfFlCMbHi\nao5vlM9vMqBXva88i4C7VNV/+l//L94HVfXDIGI40fvzjyDK/om3+sq31sDj/bL/G7AJWOxNiHnm\n4swPfh5OZ4NAzuNYtdgpwEPAXBE5X1V3+JX9g2Ofx5igWEIxsWJVEI3yTwAfAVnAZlXdX0C583B6\ndTXAqVZ6UkSWqmpagONX9P4s6gyYr+NMxZvnXzi9zOrhJIDsAvYLONe3148+jfJLRGQBzh3aIJzk\n4isTSAzyuMYAllBM2bJJVX8IolzeF+9SEVmEUzU0QUTaBGhcz5vRribOF3KohgMv+izv8jnuRuC6\nAvb7rQjnQlV3iMgunF5s/mrhzORoTNAsoRhTCFXdJSIjcRrDuwPvFlI8rx2nCc7sj6Ge6zfyTw6f\ne8+drqr+bUVFJiInAHXwq6LzdmhoSOGf1Zg/sYRiTGCv4jyX8aiIvOf3wKKvJTjVXefgtM8cJSLJ\nOL3F8tpIWohIXkP/p6p6qJDzTwduxWmIfwan0b880BS4CugWYP8853qfPclrQ3kAp03lFb9yrYBK\nwAKMCYElFGMCUNUsEXkcJ7F0488PLeaVOywiH+E8pf6s3+a7Ob595B/eFzjdl38r5PzZInIZMATo\n5y2fgdOVeTbOg5TByEtyijNn+I/AHaq6xK9cF+/2tCCPawyATQFsTDiJSArwNdBIVTe7HE6RiMga\n4H1VfcztWEzpYgnFmDATkS+Adap6t9uxhEpEuuL0NmuqqoEe+DTmOPakvDHhdw+w1Xe04VIkEbjJ\nkokpCrtDMcYYExZ2h2KMMSYsLKEYY4wJC0soxhhjwsISijHGmLCwhGKMMSYs/h/E481I/kVLzQAA\nAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Cu5PbITi45lm",
        "colab_type": "code",
        "outputId": "84210042-d194-4c80-c155-f2cfee95a3da",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "roc_auc_score(y_test_5,y_predtest_mlp)"
      ],
      "execution_count": 55,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.9976309554216182"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 55
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cJsON4305FWW",
        "colab_type": "code",
        "outputId": "344ac305-b3a4-4885-af5f-5e380b89ed01",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 309
        }
      },
      "source": [
        "plt.plot(recalls,precisions,\"y:\",label=\"SGD\")\n",
        "plt.plot(prec_forest,rec_forest,\"g:\",label=\"RandomForest\")\n",
        "plot_precision_vs_recall(prec_mlp,rec_mlp)\n",
        "plt.plot([0,rec_point],[0.9,0.9],\"r:\")\n",
        "plt.plot([rec_point],[0.9],\"ro\")\n",
        "plt.plot([0,rec_pnt_frst],[0.9,0.9],\"r:\")\n",
        "plt.plot([rec_pnt_frst],[0.9],\"ro\")\n",
        "plt.legend(loc=\"lower left\")"
      ],
      "execution_count": 56,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7f08ec3936d8>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 56
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZMAAAETCAYAAADzrOu5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nOZiYmVltDiZmZlabg4mZmdX2N9jhl0YkDaYgAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_IAr6gd_6UhT",
        "colab_type": "code",
        "outputId": "73981abe-713c-434e-b9a9-81deb53605d3",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 306
        }
      },
      "source": [
        "plt.plot(fpr,tpr,\"y:\",label=\"SGD\")\n",
        "plt.plot(fpr_forest,tpr_forest,\"g:\",label=\"RandomForest\")\n",
        "plot_roc_curve(fpr_mlp,tpr_mlp)\n",
        "plt.axis([-0.005,0.7,0.7,1.004])\n",
        "\n",
        "plt.legend(loc=\"lower right\")"
      ],
      "execution_count": 57,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7f08ec3852e8>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 57
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZsAAAEQCAYAAABm79njAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0\ndHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAAgAElEQVR4nOzdeXxcZdnw8d81M0km+96kW9qmO93S\nNtCWra0oCIqA+oiCbC4IWBQeUFzYBFn0FXwQRUBRFBE3QBEERGgLXSxtaUtbutC96ZI2zTrZZ+Z6\n/ziTkIYsM20m6STX9/MZOuec+9znmrTknnsXVcUYY4yJJldfB2CMMab/s8LGGGNM1FlhY4wxJuqs\nsDHGGBN1VtgYY4yJOitsjDHGRJ0VNsYYY6LO09sPFJEFwJXAFOAZVb2yi7Q3ArcAScDfgGtVtTF0\nbSTwW2AWsAdYoKr/6e75OTk5OnLkSABqa2tJTk4+5s/Sl2I5dojt+GM5dojt+GM5dojt+FevXl2m\nqrnHnIGq9uoL+DRwIfBL4Mku0p0DlAKTgExgEXB/m+vLgQeBROAzQCWQ293zZ86cqS0WLlyosSqW\nY1eN7fhjOXbV2I4/lmNXje34gVV6HL/7e71mo6rPAYhIMTCsi6RXAE+o6sZQ+ruBp4HviMg4YAZw\ntqrWA8+KyA04hc6j0Yx/IFFVghrE7XID0BRoIqhBvB4vALVNtShKSnwKABX1FQBkJmYCUOorxSUu\ncpOdL0N7q/bicXkYnDoYgK1HtuL1eClILwBg3cF1pMSnMDprNADL9y4nKzGL8TnjAXh9x+vkp+Qz\nadAkAP655Z+MyBjB1LypAPxpw5+YkDOBovwiAH79zq8pyi+ieEgxQQ3yxDtPUDykmOmDp9McaObJ\ntU9yytBTmJY/jQZ/A0+te4o5w+cwedBk6prrePrdpzmt4DROyj0JX5OPZ9Y/w5kjzgSgqqGKv2z8\nC/NHzWdM1hgq6iv423t/46zCsyjMLKSsroznNz3P2aPPZkTGCEp9pbyw5QU+PubjDE8fzoGaA7y4\n9UXOG3seQ9OGUlJdwsvvv8z5488nPyWfPVV7eHXbq1w44UJyk3PZWbGTf2//N5896bNkJ2WzrXwb\nr21/jc9P/jyZiZlsLtvMGzvf4ItTv0haQhobD21k4a6FXFl0JSnxKbxb+i6Ldy3mKzO+AsA7B97h\nrd1vcU3xNSR4Eli5byVL9y5lwSkL8Lg8LN+7nOUly7lx9o2ICG/tfov/lqzk0knXOn83Jct579B7\nfHnGlwFYsmcJW8u28qUZXwJg0a5F7KzYyVXTr2r9u9tbvZcri64E4NXtr1LqK+XyaZdT5mvk5a1v\nUNlQyQUTLgDgjZ1vUNvk4/zxnwLgte2v0RRoYmjdULw7jvDy+y+jGuS8cZ8A4MWtL+Jxefj4mI8D\n8PfNfyfR4+Wc0PGzm54lLT6Nj43+GAB/2fgXshOzOavwLACeWf8M+an5zB85H4Cn332a4enDW/++\nn1r3FIWZozit4HQAnlz7JBOyJzB7+GwAfvPOE0zOm8IpQ08B4Ferf8X0wdOdf3vBIE+s+TXFQ05G\nygXPtoP8ds1vmTlkFmMzJ1FVX8dfN/6daXnTGJM9lgZ/Ay9tfZHp+TMozCqkrrmOf219maL8GQxJ\nGU5FvY+FO99kYs4UBiXn4Wuq5e19K5g8aDKDkgdR3VjDqv0rmZI3ldykHKoaqlh9YDXT8ovITsyi\nsr6Sdw6+w/T86WQmZlJeX87ag2uZMXgGGd4MyuqO8G7pOoqHFJOWkMah2sNsOLS+298X3RHto+Vq\nROSHwDDtpBlNRNYB96rqn0PHOcBhIAc4M3RtYpv0PwdUVa/vIK+rgasB8vLyZv7pT38CwOfzkZKS\n0pMfq9eEE/vhxsNUNlUyNnUsAGsr13Kg4QDn5p8LwAv7X2Bv3V6+PubrAPxs288oqSvhx1N/DMD3\nN3yfQ42H+NXMXwHw7Xe/jc/v45EZjwBw49obCRLkoaKHAFiwZgFel5efTPsJAF9b/TUy4zO5f8r9\nAHx51ZcZ4h3C3ZPvxufzcd171zEmZQy3n3Q7AF9Y8QWmpE/hexO+B8D/LP8fZmXN4ubxNwNwwdIL\nmD9oPjeMvQGATyz5BB/P+ySfH/o16v3KNe9cy5lZF/CxvHMAuG3j7czLmctZeWcRCAa4c9MP+Gju\nWcwdNJemYBN3b/oh5+Sdzek5p1Pnr+O+LfdzXv65zMyYw/aqWv6y73fMzT2Tk9Im4fPX8NTuPzA/\ndx6FCYWUBxp4cf/rzMicyfCk4VQ3V7Pw0BucnHUyQxKHUtlUyeLDi5iVNYs872Der/Lxvm8D41PH\nkx6XQU1zNZtqNjEhdQJpcelUNVexpWYzE1NPIhhMoaq5nrLGw+Ql5BHviqcuUM+RpjLyvPnESxx1\ngTqONB0h3zuYOPFQG6ilvKmcwd7BeMSDL1BLRVM5Q7xDcIubGr+P2qZkXOL8/66qBAniFg8CBDXY\n4bFHnO+jQQ2i1sU7oO3+0SdXq2rxsd5/Ihc224Gvq+oroeM4oAkYBZwRuja7Tfp7gKGd5deiuLhY\nV61aBcCiRYuYN2/ecX+WaGn0N3LQd5CC9AJEhBUlK1i4ayG3nHYLixcvZk/mHp7Z8AwvX/oyAN9/\n/fs8uvpRjnz7CAAL/rWAP67/I+W3lANwzYvX8Pzm5ym9uRSAm/99M6v2r2LRlYsAeHjFw+yt3suP\nP+YUNn/a8CeqG6u5YtqXKfM18eLWF2n0N/KZkz4DwEtbX0JRPjnukzQ0B3h2w2vsPuwhJ2EUqd44\nthzZgsflYXTmaNaVVOJrPkyCx0Nuch7v7z1IgzuIx+UhOVQzqmmsxi1ukuKdNm1fYw0elwdvXCI1\nDc2U+eoQXK01rUZ/M7WNtrZfb8pMisMb50ZRUBARgPCOAaHl2PmvIKQkeMhMinMeEErvpGh5/4Gq\nykrSMzKi9Omir238bhGS4t0kxrtJ8Lg/+OidiHML8W4XcW4X8R7nT7erm5t60Dc/Ou64Cpteb0aL\ngA9Ia3Pc8r6mg2st12t6Ia4eU15fzrul7zJ72Gy8Hi+vbHuFnyz7Cc9d/BxpCWn8/O2fc/NrN1N5\nSyXp3nQW717Md1//Ll8/2amJNPobqW6sxh/043F5OHX4qShO+6iIcPXMqzkl7xNs3F8FwJWTbuXK\nSbeyencFJRV1DOGrDGn+LB/5ySIOVDWQnz4FmML8dYsAqKrPprw2lXv/+koo4iQgiQdfeCN0nAjA\nfbQcx4f+3B36043zS2Nb6FiAALC/zU/Bj9Pd1iLY7jgANLY7Dhz1c/TGuRicnkiCx/nmHe9xkZ/m\n7eIn372aBj/JCW5GZn+4M3dvyV5GFBSQkxKPN84dVn6qTpyDwojLLUJemhePu+d/kQxO97JsyVvM\nnTv3mO6Pc/dt7cb5gjinT2M4HrEc/zeP8/4TubDZCEwD/hI6ngaUquoREdkIFIpIqqrWtLn+xz6I\ns1OqytqDaxmXPY7k+GTe3P0mC/61gD9/9s9MzJ3Iv7f/my88+wXWX7ueyYMm4w/68TX5qGqoIi0h\njXPGnENmYiYel/PXtOCUBVw2+Rqq6tyU1Qc5bdzFpAbO4ba/byIrOY6XNyRzxDePjZuWsn5fFcHW\nL/1Lwop3Z1ltl9eHpHf9izItMY60xDjW7Kngi7NHkJUUf9T1I7VNTC9wvtVt2rSJebOmtxYQ4chJ\nSSAp/uhf7glxblISevef8aJFh5g3b2L3CU9QHpf0eaFhBp6+GPrsCT3XDbhFxAv4VdXfLunvgSdF\n5Gmcr8K3Ak8CqOpWEVkL3CEitwLnAlNxBgj0maAG2V25m6S4JPJS8liyZwlnPnkmz33uOS6aeBFp\nCWmMyhzV2qQwe8iZvHDxa2R7h1Hd0MyZBefw7y+egz+gvLBuP3GubAZ7zuXNLdX8esk6Vu+uOPqB\ni9/qMI51JVUfOjdx8NEVwX0VdYwelEIwqBSPzOLU0dkUZCV9qFoe53YxON2Lp4d/OaVXvs/swuwe\nzdMY0zVVxe8vx+1OQdXP5s1fIjV1BgUFtwCwZ8+PARcFBU4/6e7d9+J2JzNs2PHWa/qmZnMrcEeb\n4y8CPxCR3wDvASep6h5VfUVEfgwsxGmvebbdfZ/HKXwqcObZfFZVD/dC/K0CwQAvbn2R/JR8Zg2b\nRXVjNWMemsYNxXczJfNTlNflMj/tOf66JJf/rHkHgEH+27jwod00+Xe2yWlpxM/O9gper5eq+mZO\nGZVFUJWZBZlUNzQzd9wgUrwePC6hMDeZpPgTuQJrzMCmqjQ07MDjySYuLoO6ui00NOwiJWUm8fE5\nNDcfoaZmFampJxMXl0VTUxk+32pSU2cRF5eBz/cuDQ27yMlxRu4dPvwstbUbGTnydsrL/8PWrdeQ\nl/dFRo26k6qqN1m7dh6TJv2NtLQ5JCaOobT0j62FTXX1fxH54PdFVdUyPJ6e6SPri6HPdwJ3dnL5\nqOFVqvogzlyajvLZBczrucjCs2jXIgDmjZzHEV8z1/75ZbyN5xIMvgTAcJ7h2SXwLO+F7ohnR2kV\n8OHaRovUDpqBahr9pHo9nDY6p/VcnMfFvRdNJtUbd8IPbjAm1jU3VyDixuNxWgWqq98mPn4IXu8w\nVJXKyjdISCggKWksdXVb2bHjFiZOfBq3O4kjR16mqamUwYOvBGDPnh/h860FvgbArl0/oLJyMdOm\n/Yfa2vdYtWoKY8Y8xLBh36C8/DVKSh5k3LjHyMr6GD7fu7z77scpKlpERsZcfL7VvPvux5k+fRnp\n6XOoqHiDkpKfUl+/g+HDb6CycjFHjvyTkSNvJy1tFkOGXEMw2ABAWtps8vO/THLyNBIShjBq1N0U\nFt7T+pknT37uqJ/B1Kkv9tjP077yhiEQDOB2uWkOBPnOqw9SuvuLaKhwied8gp3cNzI7idG5KeSk\nJFA8MrO1MzkQVDKS4igankFGu34NY8wHGhv3oerH6x0ROvN3qquTSEs7haamw+zZ8yPy8i4hNXUG\njY0H2Lv3AfLzLyclZSpVVcvYu/cnjBnzf3i9BVRWLqaiYiEFBd/B7fayefOX8fneobh4DQCbNl1G\nXd1WZs5cAcCaNaeTnHwSkyb9FYB33pnDiBHfY9Sou1ENsG7dRxk58m5GjPg+dXWbKSv7OzU1K8nI\nmEtDw0727LmPYLCWoUO/jmqQo3sKBK93FPX1O0hJmcyIEbeSmXk2AEOGfJXU1JkkJTn9gqmpM5g+\nfRnJySeFjmeFjie3ps/L+wJxcc58trFjf8bYsT8DwONJbW0SA3C5Epgw4dcfRCG913dnhU0X6pr8\nPPCf//KbJQfQYFLo7Nc+lG5oRiITB6dx/2emkJUUj6sXhyMac6ILBv0EAlXExTl9dLW1Tq2/5Zfn\nzp13kJV1Lunps6mv38WGDZ9i8uQXSEwcydat19DYuJ/i4tWh3H7GkSNppKWdgt9fyf79j5KWdgqp\nqTPw+8vZv/9R0tNPJyVlKoGAj/r6bQSDTQD4/TWUlT2LyxXHiBHfJyNjHomJY1rjzMg4i+TkKa3H\nI0Z8H48ns/V4ypR/tqYXcTFlykskJAxHRMjO/iRnntmIy+V8eRw69DqGDr2uTV7fBZzRaAAjR95+\n1M9o1Ki7W9+7XAmkp38wYs3jST/qOC4u46hjtzsZt/vEXwLHCpsO1DQ0M+XOf7c5k3TU9cQ4N/XN\nAZZ/9yPkp3lb5xIY0181NZXi91eTlORMEK6ufpumpgPk5FxAY+N+3n//6wwbdiPp6WewefPl1NZu\nZOLEp0lOnsimTV/kyJEXOfNMHwBbt16LiIuiooUA7N59F42Ne0hPn00gUEVzcxk1NStITBzJ8OE3\nEwjUtYnkWYYPd1YBSEoa25onQHLypKOOs7LOJivr3dbjnJxPkpPzydbj/PzLjvqMLU1eLfLyLjnq\nODv7vNb3Iq4PHYtYK0VXrLBp5+X1B7j26XeOOpeS4GHO6Gxu/+RJDMtMtMLFnJBqazcRCPhISzuZ\npqZDHDjwG7ze4eTlXQrA1q0LSE8/FRgCwJYt15CZ+REGDfocTU2lbNz4OQoKbiE7+zwaGkrYtOlS\nRoz4PllZZ7N7972Ulv6e00+vCOX1NZqaDpKVdR7BYBP19TsIBHyICOPGPca2bTe0fssfMuRqvN6C\n1jgLC3901P9Ds2ZtJyHBuZ6SMo1TT/1gHlZGRvv5QJl4PLG56sdAZ4VNyL83HuSmv6yjpvGDdtUU\nr5+1t53f48N+jQmXaoD6+m14vaNxuTzU1m6kpmYV+flXUFGxkPffv54xYx4gNbWY3bvvprz8FU4/\nvRwQGhv3Uln5emthU1W1ONSU5RQ2lZUL8XqHtz6rbfu9iISOnUIhP/9y0tPPaL0+btzj+P0VuFxx\nJCaO5OST17Vec7uTGD/+8dbjzMyPkJn5kdbj9PTWhT8ASEwsPO6fkznxWWEDfO2pVby6sfSoc7On\nrOaGuWdbQWOiwu/34XLF43LF09i4j4qK18nPvxyAxYvjmTDhd+TlfYG9e3/Kjh3f4tRTDxEfn8u+\nfb+krOx50tJmk5AwlKysjxMI+IiLy2bMmJ/i8zkLYcbH57Z2Erc4+WRnMcXduxcBMGvWltZr8fF5\nrc1aAAkJQ486Tk2dSWrqzNbjtLSTe/YHYvq9AV/YXPuH1UcVNF87s5BvfnQsSfGf6MOoTCxqbi7H\n7U7F5Ypj//5f4/dXUlBwM9XVK9i27SYGDbqYYcOup6pqGWvWnMbUqa+SlXU2+/f/it27f0BW1rnE\nx+eSnn4Gzc3OlLHc3Ivw+4/gdqcCMHz4zQwefFVrTWfMmJ+0Pj8+Po+srLzWY5HwltIxpjcM6K/t\nu6oCvLzhYOvx+/ecy6knHWJn5ZYu7jLGsX//42zdugCA2tqNLF2aTUmJswJ2RcV/OHLkBQBSUmbQ\n2FhCS5NUWtosUlJmEB/vNGfl51/J1KmvtI7WKip6nWHDvgFAYuJoCgvvw+32ho5Hkpo6E5drwH9P\nNDFmQP+LvXN5Q+v7dXecTZzbxTde+QZucbP66tU2EKAfUw0g4qap6RDV1W+Tnn4acXGZHD78LDU1\nqyksvBeAnTtvIxhsYvToHwGwdu1ZwFCCwdPw+d5l//5fMHLkbSQlnUR29vkkJo4CYNKkP7U+y+WK\nY86cXa3HIu42Q3mdAiQxcWTUP7MxfWnA1mwOVn1Q0Dz8hemkJzpLnC+8YiFPf/ppK2himKpSU7MW\nv98ZBrt9+7fZvPnLrdeXLMlk27YbQ+lWs2HD+dTXbwWgunoFhw4905q2qelwa5MWODUNyMblimP0\n6B8xd26Q+Pg8RIQpU14gN7dPl+cz5oQ1YAubp/67q/X9+dOGtL7PSsxiYm7srug7UAUCDezefS/N\nzeX4fOtYvXo6O3Z8G9UgInG4XAmtaYcMua51Ml56+qnMmLGSpCRn98/Ro3/M7NkfrFs3fvyjTJjw\nmzbHjwPObpJud7J9KTEmTAO2sCmrcWYVj8h2JmzuqtzFRX++iDUH1vRlWKYLqkp9/Q4aG51+tn37\nHmXzZmf0VVPTfnbu/D5VVctITS1i2LCbyM39HCIuCgvvYdy4R1rzKSy8h4KCbwHO7Oy0tGKbu2FM\nlA3YwmZDaEOxT08fBsCBmgNsOLShL0MyIQ0Ne0JrYimbNl3Ghg0XhdaWambFitEcOtSybZFy8OBv\nCQb9eL2jGDPmIbKynD3nx4z5CZmZ8/rsMxhjjjZgC5t9lfUAjB7krCk0Z/gc3r/+faYPnt6XYQ1I\ngUAdW7ZcQyBQS3n5f/jvf0dw8OBTQJCkpJOorl6Jqh8RN+PGPd463yM//wrmzvXjcnkQEYYN+4aN\n0jLmBDVgC5uM0ICAUTlOYaNqe9n3lqamUuB+Nm78PAD19Ts4cOAxDh/+G1lZH2XatNdJSBiCiJsR\nI77LqaeW4HLFI+JmyJCvti5h4nYn2VwSY2LEgC1sdh1xFvfLTUlg0+FNjH14LMv2LuvjqPq3rVuv\nY9++X+Bs0ppKfLwzATExcRTTpr3OoEHOwoeZmR9pnU1vjOkfBmxhE+d2RhGleuOoa65jQs4EhqcN\n7+YuEw6/v4ZgsBGAxYsT2LnT2WA1Li6b999fAASArzN2rDMB0u1OJjPzI7hccX0UsTEm2gZkYaOq\nNAecZrMEj4uZQ2by4iUvMjzdCptjoRrgwIEn8Ptr8PurWbIkjR07vgPA8OHfIj39dMDZs2PePG2t\n0RhjBo4BWdg0BZy9Nd0CLpcQCAb6OKLYVFW1HNUA5eWvsmXLV6isXIjHk0Ze3hdbdw0sLPwhWVkf\n6+NIjTF9LeyhOyIyG/g4MBtnjfJEoAzYAiwG/q6qFdEIsqc1+kOFjQvqmuvI+XEOD57zINcUX9PH\nkcWOmpo1bNjwKbKzP8WECU8wePBXW5egnzjxqT6Ozhhzoum2ZiMiV4jIemAZcCPOtpXvAyuACmAW\n8Gtgn4g8KSKjohhvj2gOFTZNAWgKNHFN8TVMGTSlm7sGrurqVQSDjdTVbWXlyiL27v0pSUnjSEub\nTUHBLYAzsz4uLrObnIwxA1WXNRsReRfIBX4PXA6s1Q7GCItIOvBJ4FLgPRG5UlX/HIV4e0Qw9AlS\n4yHDm8GD5zzYtwGdgAKBWsDN7t0/ZM+ee5g06Tlycy8iOdlZ1sXlSmLKlH/2bZDGmJjRXTPaE8Bj\nqtrQVSJVrQKeBp4WkWlAfg/FFxXBUHnpEqE50IwnNClwIAsE6ikvf5nc3E9TW7uZlSsnMnny3xk1\n6m5AiY93/kpPOunpvg3UGBOTumxGU9WHuitoOrhnnaq+enxhRVdrYQP8bt3vSL43mX3V+/o2qD7Q\n0LCHgwd/B0Bp6dNs3PgZGhp2k5w8gYyM+cTFOasZFxbeQ3r6nD6O1hgTywbk2h6BUDuaCEweNJlr\ni68lL2XgDMdtaCjB6x3G4cN/Y/v2m0hLm0N29rkMG3YTcXGDACgqeqOPozTG9Cfd9dncHkFeqqp3\nH2c8vaKl18klMHvYbGYPm923AfWiTZsuo7T0DxQW/pjc3M/i8aTj9Y7A5Uo4aothY4zpSd3VbO6M\nIC8FwipsRCQLpz/obJzh099V1T92kC4DeAg4N3TqEVW9s831XUAezpR0gGWqenZ3z2+t2QA1jTWk\nxKf06z6b+vqdrFhRyLRpCxk37lECAR8ZGfNJTBxFYuKXu8/AGGOOU3d9Nq4IXpGsiPgLoAmnoLgU\n+KWITOog3U9xhlqPBE4BLhORq9qlOV9VU0KvbgsagEDrAAE4+Vcnc8lzl0QQemwIBptZsWIcPt+7\nBAI+cnI+zYYNn8LtTmby5OdJSyvu6xCNMQNIr/fZiEgy8Blgsqr6gCUi8gJwGfCddsnPB85V1Tpg\nl4g8AXwJ+O3xxNAyelsEbph9A0NSh3RzR+woKfk5SUnjyciYh4ib3bvvZdKkPzFp0l8RGZALRhhj\nTgDS20vri8h0YKmqJrU5dzMwV1XPb5e2DDhPVd8OHX8fuFlVM0PHu3BWMnABa4Bvqeq6Dp55NXA1\nQF5e3sz7H3uaO5c3kJ+o3D83Nndo9Pl8pKS0jb0JiMeZDtUInLDTnICO4o8dsRw7xHb8sRw7xHb8\n8+fPX62qx94koqphv3B+Ya8B6nD6SY56hZnHGcDBdue+CizqIO0fgOeAVGAMsB1obHP9NJzCJgn4\nLnAQyOjq+TNnztT1JZU64pYXdcQtL+rh2sPqD/g11ixcuLD1/cqVRbpwIdrQcECrqt7WxsaDfRdY\nmNrGH2tiOXbV2I4/lmNXje34gVUaQXnR/hV2u4qIXA48DKwEvDhNWX8AqkOFwF1hZuUD0tqdSwNq\nOkj7DaAeZ3mcfwDPACUtF1V1qarWq2qdqt4HVOIUZmHJT24m9//l8o8t/wj3lhNKWdkLNDdXMHHi\n0yQkDKexcS9paSfbqsrGmBNOJI34NwD3AdeGjh9R1SuAQpwC4UiY+WwFPCIyts25acDG9glVtVxV\nL1XVfFWdFIr37S7yVpxBZl1qaTl0i5uHPv4Q0/NjcSvocjZsuIClS7NITBzDnDl7SEs7ua+DMsaY\nDkUyQGAs8CYQDL3iAVS1QkTuAe4Bft5dJqpaKyLPAXeJyFeAIuAC4NT2aUVkNE5tpRJnmPTVwNzQ\ntQJgOE5NywVcD+QAS8P9QB6Xh2/M+ka4yU8ITU2lrFo1E3iSWbO2U1+/A5crvq/DMsaYLkVSs6kH\nXKG2u4M4NZoWPpxtB8J1HU5fyyGcprFrVXWjiJwhIr426WYC63Ga2O4DLlXVlhpQKvBLnJWn9+Fs\nf3CuqoZbwyKoQQ76DhLUYASh961Dh/5EU9M+4K8kJhaSlfXRvg7JGGO6FUnNZj1OJ/1/gLeA74nI\nTsCPM/lzc7gZqWo5cGEH598CUtoc/wX4Syd5bASmhh9+m3tx2tGqm6sZ/MBgDt508IRerqaqaikb\nN36OnJwLKSy8l8zMc1i58kBfh2WMMWGLpGbzONCyYcltOIXCEuC/wDjgpp4NLfqSPIk8ct4jZHgz\n+jqUTjU0lJCWNptJk/5GQsJw3O5UkpMnEEbXlDHGnDDCrtlom/1pVHVbaMb/HJxhx8tUtSwK8UVV\ngiuBa0++tvuEfaS2diMrV05m2LAbGDPmp7bysjEmZh3zCgKqWovTpBZzWkajBdTPnqo9FKQX9G1A\nHaitfQ+vdySjRz9AIODr/gZjjDmBRVzYiEg+UIAz1+YoqvpmTwTVW440HWHao9OouKWir0M5SlnZ\nP9iw4UJOPvk9hg//374OxxhjjlvYhY2IDAWeIjT0uP1lnDkukSzG2efSPGncet4v+jqMVnv3/h/Z\n2eeRkTGPhIQRxMVl93VIxhjTIyKp2fwSmAJ8G2dkWmNUIuoFLavBJboTuWTKp/s0lhZ+fxW7dt3G\n9u03MmdOCXPm7OrrkIwxpsdEUticAXxDVZ+KVjC9rUmbOFBzgMGpg/s0jv37H2fQoIspLl6PiJCQ\nMLRP4zHGmJ4W6aTOQ9EKpC/sr9/PVf9ovz1O71INcOjQn1myJAOvtwCvd0SfxmOMMdEQSWHzK5w9\nZ2Kehoaj5Sbk8t3Tv9uncagGGTbsRoqK3rL9Zowx/VYkzWj7cHbKfB14GShvn0BVf9NTgfWGFE8y\nc0d2NN6hd6xcOYmUlBmMH1/rtDMAACAASURBVP8Ybndyn8VhjDHRFklh82joz5HA/A6uKxBThU19\noIFSX2mfLFWjGiA//ypKSn5qBY0xpt+LpLAZFbUoelnLaLRdtTv5xco13DU/3K14esaePf+PxsYS\nRo9+gOHDb+7VZxtjTF+IZLma3dEMpC8MTxrGFyaf2+vPjYvLYseObzNq1D14PLG5RawxxkTimJer\n6Q9SPKlMzJ3Ya88rKXkItzuV3NyLiYvLsYLGGDNgdFnYiMgO4CJVXRfaTkC7SK6qOrpHo4uSlrXR\n6gP1lNWVkZOU0yvP3bv3QYLBBvLyLiMn54JeeaYxxpwIuqvZLAaq27zvqrCJObtqd/La9ka+MOUL\nUX9WMOhn9uwdNDTsxeWKi/rzjDHmRNJlYaOqV7V5f2XUo+llBUkFzB15WtSfU1+/nc2br2LatNdI\nTBwZ9ecZY8yJ5rhnEYpIDK4W6VTQUjwpDEmNZDfrY1Ne/hpVVW9RU7Mm6s8yxpgTUdiFjYh8VUS+\n1eZ4ioiUAIdEZFVo64GYUheopaqhKqrPaG4uZ9Cgi5k1azvp6bOj+ixjjDlRRVKzuR5nfbQWDwKV\nwA1AOtC7k1V6wM66Xbx3+L2o5a8aZOnSbI4ceYnExMKoPccYY050kRQ2I4DNACKSjrOvzbdV9WHg\nDuCcng8vOlpGo41KGslJuSdF8Tl+hg//FmVlz0btGcYYEwsimWfjAoKh96fjdHwsCh3vBQb1XFi9\nI9mTTLo3PSp5B4ONHD78HCNG3IbbbfNpjDEDWyQ1m/eBT4Tefx5Ypqp1oeMhdLAw54nO5/dR31zf\nfcJj8N57l7Bt2zfw+ysQkag8wxhjYkUkhc1PgBtEpAy4BHi4zbX5wLs9GVg0ta6NVreLw3WHo/KM\nESNuJT39TBIShkclf2OMiSWRrI32RxHZA8wCVqrqm20ulwIv9HRw0VaYXEhecs+v+NzcXE5y8mQm\nT7a+GmOMgQjXRlPVJcCSDs7f0WMR9aJkdxIJnoQez3fp0mxSU4uZOXNlj+dtjDGxqE+2hhSRLBF5\nXkRqRWS3iFzSSboMEfmdiBwKve5sd32kiCwUkToR2SwiHw3n+S2j0Wr8NfiD/uP8NO3zViZOfIa4\nuJgbL2GMMVHTV6s+/wJoAvKAIuAlEVmnqhvbpfspkISzYdsg4HUR2a2qvw1dfwZYDpwXev1NRMaq\nalgdMbvqdtEcaMbj6rkfg99fwaBBF5OX9/key9MYY2Jdr9dsRCQZ+Axwm6r6Qk1zLwCXdZD8fODH\nqlqnqruAJ4AvhfIZB8wA7lDVelV9FlgfyrtLGqrajEkZ06PNaH5/FUuXZnPgwOM9lqcxxvQHfVGz\nGQf4VXVrm3PrcCaJdkTavZ8cej8J2KGqNe3ymfShDESuBq4GyMvLY+3atQAkEM+bi99sn/w4lAOn\ns3Xrm2zdOr4H8/0wn8/HokWLovqMaIrl+GM5dojt+GM5doj9+I9HXxQ2KXywbUGLKiC1g7SvAN8R\nkStwmty+hNOs1pJP+4XNqoCh7TNR1ceBxwGKi4u1qKgI3v4vvqCPefM+ecwfpK26um243RNISPh0\nj+TXnUWLFjFv3rxeeVY0xHL8sRw7xHb8sRw7xH78x6MvBgj4gLR259KAmg7SfgNnPbb3gX/g9NGU\nHEM+R2mZZ7Onbk94EYdh69avsnz5MPz+bh9vjDEDTkSFjYhcLSJrQqO/Au1fYWazFfCIyNg256YB\n7QcHoKrlqnqpquar6qRQvG+HLm8ECkUktbt8OjM2ZWz3icI0YsStFBR8G4+nowqaMcYMbJFsMXA5\nzqoBKwEv8FvgDzhNYtsJc9VnVa0FngPuEpFkETkNuAB4qoNnjhaRbBFxi8i5OP0uPwzlsxVYC9wh\nIl4RuQiYCoQ9kzLR7Q03abcyM8+isPDeHsvPGGP6k0hqNjcA9wHXho4fUdUrgEKcpq4jEeR1HZAI\nHMJpGrtWVTeKyBki4muTbibOCLOa0LMvbTc8+vNAMVAB3A98Npxhzy3zbKqae2Yvm0WLhE2bLmsd\n5WaMMeZokQwQGAu8ibPycxCIB1DVChG5B7gH+Hk4GalqOXBhB+ffwun4bzn+C/CXLvLZBcwL9wO0\nV1K/91hvbRsD48c/QUnJQ7bgpjHGdCKSmk094FLn6/tBnBpNCx/Oys8xZUwP9dkMHvwlTj55XY/k\nZYwx/VEkhc16YEzo/VvA90RkjoicDNxJaGO1WKCh8WgJruOb0Flb+x6LF7soLX2mJ8Iyxph+K5Jm\ntMf5oDZzG/AfPliUs4YOmsVOdFXNlcd1v2qA/PyrSEj40NQeY4wxbUSyxcCf27zfJiKTgDk4kyyX\nqWpZFOKLqn0N+4/r/pSUKYwd+wjuHhzVZowx/dExT+pU1VpV/Y+qvhBzBU1o0NjxzLMpK/sHK1dO\npa7uvR4Kyhhj+q8uazYiUgAcUNXm0PsuqWrPTcnvBQmuuGO+NzW1GL+/0prQjDEmDN01o+3EaSp7\nG9jFByu9dMbdAzH1morj6LOJjx9MUdEi4uN7fqdPY4zpb7orbL6EszpAy/t+MWux5UOUNhw8pvsP\nH/47ZWXPMn78b3ouKGOM6ce6LGxU9Xdt3j8Z9Wh62ZiUMd0n6oDLFU9p6R8oLLzfmtGMMSYMkayN\nlhvasKyja+NEJKfnwuodHjm2HRays89j/PjfWEFjjDFhimQ02iPATZ1cuzF0PSa0LGFWeQx9NmVl\nL1Jfv5PBg6/q4aiMMab/iqSwOR14tZNr/wZOO/5wetehxkMR37N//y9YsaIQ1WAUIjLGmP4pknak\nTD68M2aLaiD7+MPpXWNSRkd8z7hxj1FZ+SYifbHvnDHGxKZIfmOWALM6uTYLOHD84fSOlrXR3BL5\nSG2vt4D8/C/2dEjGGNOvRVLY/A34roh8ou3J0PF36GIrgBNVZVNkfTYNDXvYsuUa6uq2RCkiY4zp\nnyIpbO7CWfn5BRHZJyJvi8g+4IXQ+R9EI8BoOtzY7T5rR/H7K6muXk59/bYoRWSMMf1TJAtx1onI\nXOAy4GM4fTTbcAYH/EFV/dEJsee1jEYbHWGfTUrKVAoKvk1GxvwoRGWMMf1XRBNNVLUZ+E3oFfPc\nEXby19dvJy/v0ihFY4wx/deAHlJV3lQRdtqamrWsWDGGkpKHoxiRMcb0T5GsIBAvIneIyGYRqROR\nQLtX7DSjhf480nQk7Hs8nlQGD/4q6emnRicoY4zpxyJpRvt/wNeBl4HngMaoRNSLIpln4/WOYty4\nxxCRKEZkjDH9UySFzWeBO1T1nmgF09tcEbQi7tlzPxkZ80lPnxPFiIwxpn+KpM8mBVgerUB6k4aG\no0XSjLZr1x1s23ZjtEIyxph+LZKazT+BM4E3ohRLr6toLg877amnHiYYrI1iNMYY039FUtg8DPxe\nRILAv4AP/aZW1R09FVhvGJMc3n42waAftzuRuLiMKEdkjDH9UyTNaMuBscCdwArg/Q5e3RKRLBF5\nXkRqRWS3iFzSSboEEXlUREpFpFxE/ikiQ9tcXyQiDSLiC73CXkOmdbvRMDv7S0t/z8aNFxMMxvyY\nCGOM6ROR1Gx6alvoXwBNQB5QBLwkIutUdWO7dN8E5gBTcVabfhyndvXpNmkWqOqvjzWQI41lYaUL\nBOqorl6GSPyxPsoYYwa0SJarefJ4HyYiycBngMmq6gOWiMgLOEvgfKdd8lHAq6paGrr3z8CDxxsD\nQHV9MxD+5mnDhi1g2LAFPfFoY4wZkCJeQUBEXCIyWUTmhgqPSIwD/Kq6tc25dcCkDtI+AZwmIkNE\nJAm4FGeOT1v3iUiZiCwVkXnhBpEc75SxXi0MK31p6Z/CzdoYY0wHIlobTUS+DtwB5OA0qZ0MvCMi\nfwfeUNWfdZNFCs5Ga21VAakdpH0f2AvsAwI4K0u3rV7cAryH0yT3eeCfIlKkqts7iPtq4GqAvLw8\n1m9YD8DQpCCLFi3qJuQ64Ats2rQfmNFN2t7j8/nCiP3EFcvxx3LsENvxx3LsEPvxHxdV7fSFU/O5\nOfT+q4Afp+/ks0AQmBG6dhOwuKu8QummA3Xtzt0E/LODtH8AngeygATgNmBFF3m/AlzfXQwzZ87U\n1zYe1BG3vKin3vN77U5Dw37dsuUaPXTo2W7T9qaFCxf2dQjHJZbjj+XYVWM7/liOXTW24wdWaTe/\nX7t6ddqMJiIzgFV88HX+f4EHVPXqUCHQ1mZgfBhl21bAIyJj25ybBrQfHADO4IEnVbVcVRtxBgec\nIiI5neStQERryVQ1d7bL9QcSEgYzduwvyM39dLdpjTHGdKyrPpsVwFOq2jI0eRTwaidpa4FuJ6Go\nai3Oump3iUiyiJwGXAA81UHylcDlIpIuInHAdcB+VS0TkQwROUdEvCLiEZFLcSacvtJdDG2NTul+\nnk1NzVoaGvZEkq0xxph2uipsfgPcENowDaAMGNlJ2vE4fSvhuA5IBA4BzwDXqupGETlDRHxt0t0M\nNOD03RwGzgMuCl2LA34YOl8GXA9cqEcPPOhUJOO3t2+/mRUrRkVwhzHGmPY6HSCgql8TkTOAR4Ap\nwIvA7SKyCNjdkizUrHUj8PdwHqiq5cCFHZx/C2cAQcvxEZwRaB3lcRhncMJxKWs81G2awsL78fnW\nHu+jjDFmQOtyNJqqviUiM0OHtwLzgQ04TWwK/AyYgFNLuSuKcUZFjd/XbZq0tGLS0op7IRpjjOm/\nup1no6pNoT/LgGLgPpxmrO04hdXPgTmq2n1v+wmmMLnreTbBYBN79/6f9dkYY8xximiejarWAHeH\nXv1eXd0mtm+/kWCwgREj2i9wYIwxJlwRFTb9zaFu+mySkk5ixoy38XoLeikiY4zpn7osbETkDeA6\nVd0cet8VVdWzei606NHQ5ml1/q73pwkEfKSlHfc4BGOMGfC667NpO0nSFTru7BXxOmt9bVRK10Oa\n1637KLt2xdy4B2OMOeF0Nxptfpv386IezQnG48nE72+/lJsxxphIDew+m4au+2ymTXuVGKywGWPM\nCSfs36QicouIPNzJtZ+JyLd6LqzeUR+o7yaFIGHu5mmMMaZzkXxtvwp4t5Nra0PXY8qI5BGdXtu7\n90HeeisNn299L0ZkjDH9UySFTQHOOmUd2QF0/pv7BBPO2mhpabNISZlGYuLY7hMbY4zpUiR9NnXA\n0E6uDQMajz+c3nWoobTTa+nppzFjxtJejMYYY/qvSGo2bwHfEpGEtidDxzeFrseUxmDn5WNZ2T9o\nairrxWiMMab/iqRmcyewDNgqIn/A2VJgKPBFIBu4sqeDi7bhSR2vDBAMNrFhw4Xk5V3BxIlP9m5Q\nxhjTD4Vd2KjqOhGZD/wEuAWnVhQElgCfUdV10Qmxb0yZ8jJxcVl9HYYxxvQLkS7E+TZwpogkAplA\nhap2N374hFXaSZ+NyxVPdvbHezkaY4zpv45pxqKq1qvq/lgtaEJLo+HX5g6vV1a+yb59j+D3x9yu\nCcYYc0LqbiHO24Ffq+r+0PuuqKrG1NYDwxKHdXi+pmYV27ffRG7u//RyRMYY0z9114x2J/AKsD/0\nvitKP9nnZujQb5CffwUej/XZGGNMT+huIU5XR+/7i873swkSF5fdq7EYY0x/1mUBIiLviMik0Pvb\nRWRI74TVOwIa6PD8tm3fZO/e/+vlaIwxpv/qrrYyBUgJvb8DZ6WAfmOwd3CH5ysqXqeq6s1ejsYY\nY/qv7vps9gMXikgpzgZp+SLS6R7JqrqnJ4OLnq5XRzvllM00N5f3UizGGNP/dVfYPAb8EPg2zm/o\n57tJ7+6JoHrLocaO59mIuIiPz+nlaIwxpv/qboDAvSLyGnAS8FvgPpwVnmNayzwb7aCGU1u7mc2b\nL2f06AfJyDi9lyMzxpj+qdsVBFR1JbBSRK4EnlLVzVGPKspa9kPL9+Z/6Fog4ANctmmaMcb0oLCH\nM6vq/J4oaEQkS0SeF5FaEdktIpd0ki5BRB4VkVIRKReRf4rI0EjziVRaWjFTp75EauqsnsjOGGMM\nES5XIyLTReQ5ESkTEb+IzAidv1dEwl1M7BdAE5AHXAr8smV4dTvfBOYAU4EhQAXQdlvqcPP5kEAw\nCHS8NprfX41IPC5XRMvGGWOM6ULYhY2InA4sByYAf2x3bxC4Jow8koHPALepqk9VlwAvAJd1kHwU\n8KqqlqpqA/BnYNIx5PMhH/TVfLiprKzsef773xE0Nh4IJytjjDFhiOTr+/3Aq8CFOKPOFrS59g5w\neRh5jAP8qrq1zbl1wNwO0j4BPBSaSFqJU3t5+RjyQUSuBq4GyMvLY9N7TmtgtieLRYsWtUt9CJjG\n8uVrgC1hfKS+4fP5Oog9dsRy/LEcO8R2/LEcO8R+/McjksJmBvBpVVURaT+MqwzIDSOPFKC63bkq\nILWDtO8De3E2aQsA6/mggIskH1T1ceBxgOLiYp08aRKsfQePx8O8efPapZ4HfCuMj9K3Fi1a1EHs\nsSOW44/l2CG244/l2CH24z8ekfTZNABJnVwbjPPLvjs+IK3duTSgpoO0vwAScHYBTQae44OaTST5\nfEhz0Nla4FAn+9kYY4zpWZEUNkuAG0Sk7cTNlhrOl4E3wshjK+ARkbFtzk0DNnaQtgh4UlXLVbUR\nZ3DAKSKSE2E+nRL58Md/770vsnLl1EiyMcYY041ImtFuA5bi9I38DaeguUJEHgRmAid3l4Gq1orI\nc8BdIvIVnALlAuDUDpKvBC4XkUVAHXAdsF9VywAiyOdD4lxxAOQmfLjlLzv7PJzxDsYYY3pKJPNs\n1gFnAqXA93GGcrX0ocxV1XB7068DEnF64p8BrlXVjSJyhoj42qS7Gafp7n3gMHAecFF3+YT1Wbq4\nlpd3CSed9McwP4oxxphwRDSZRFXfAc4SES+QBVSqal2EeZTjjGhrf/4tPlhhGlU9gjMCLaJ8wtHk\nbwI67rMJBptxhWo+xhhjesYxbYimqg2quj/SguaEEZpe4+qgz+bNN+N5//0bejkgY4zp37rbPO1/\nQ7WYsInIjAhWE+gT8e544MN9NqrKqFE/JDV1Rl+EZYwx/VZ3NZvLgJ0icr+ITOsskYhkishlIvJv\nnFFr7YclxwQRYcSI75OfH878VGOMMeHqrs9mBk6BcxPwbRGpxplceRhoBDKBQmB06PjPwEmquita\nAfeERn8jAAfb9dk0N5fT2LiXpKRJtjaaMcb0oC5rNur4vapOw1kU86c4EycLgek4M/bfAr4EDFHV\nq070gsbhdNp4XEfv9VZTs5pVq4ooL3+lL4Iyxph+K+yv76q6AlgRxVh6TUufTU673TiTk09iwoQn\nSUuz7QWMMaYnWVtRGwkJQ8nPv6KvwzDGmH7nmIY+x7pGfwMABxuO3kagqakUn289qoG+CMsYY/qt\nAVnYuFzOx453xR91fteuH7Bq1VSCwaa+CMsYY/qtAdmMFhcqZLLis486P2zYjaSkTMftTuyLsIwx\npt8akIWNdrI6WlLSWJKSxnZ4zRhjzLHrsWY0EUnoqbyi7YN5Nkf32VRWLqamZk1fhGSMMf3acRc2\nIuIVkRuBHT0QT6+Q0DybBNfR5eO6deewe/ddfRGSMcb0a902o4lIAfB5oADYBvxWVatEJB64HmcP\n5UHA8mgG2pPiPaE+m4Sso86fcspGGhv39UVIxhjTr3VZ2IjIacBLHL3W2ddE5FM42zRPAt4BvqSq\n/4palL0kMXE0iYmj+zoMY4zpd7prRrsDOAKcDiThFC6lwDKcJWuuUtXiWCto6pvrAdhfv7/1XDDY\nyO7d91NXF+4ecMYYY8LVXWFzMnCnqi4L7WGzCWeHzGzgO6r6u6hHGAUt+9gkthni3NRUys6d3+Xw\n4ef6KixjjOm3uuuzScfZlrmtluO3ez6c3hHvdgYGZMVntp5LSBjK7Nm78XgyO7vNGGPMMQpnnk37\ntVuCoT/71TR7ETdeb0Ffh2GMMf1SOIXND0SkrM1xaFNl7haR8jbnVVVjYhXL2qZaAPa16bOprl5B\neflrDBlyDfHtVoM2xhhzfLorbPYAEzs4vxtnsEBbHU/LPwF5QhujJbmTWs/V1m5g167bGDTof6yw\nMcaYHtZlYaOqI3spjl6V4HH6bDLjM1rP5ed/iUGDLsXVbnFOY4wxx2+Aro32YSKC2+3t9ViMMWYg\n6Ha5GhG5UkTWiohPREpE5IHQ6gExq6axBjh6ns2BA0+wc+ftfRWSMcb0a10WNiJyCfAbnAmdLwH7\ngRuAe6MfWvTEueMASPYkt56rrl7B3r0P9FVIxhjTr3VXs/km8DwwUVUvVtVTgLuAr4uIO+rRRYnX\n4zSXpcelt54bP/5xTj+9sq9CMsaYfq27wmYc8Cs9ep/kR4AEnIU5j4mIZInI8yJSKyK7QzWojtK9\nHGq+a3k1icj6Ntd3iUh9m+v/PtaYAFyuuOO53RhjTCfCWUGgvN25luNMYOcxPvcXOJNC84Ai4CUR\nWaeqG9smUtVz2x6LyCLgjXZ5na+q/4nk4VUNTg2mbZ/Ne+9dSlraKQwb9s1IsjLGhKG5uZmSkhIa\nGhqOK5/09HQ2bdrUQ1H1vliI3+v1MmzYMOLievbLdzij0Vwi0rYG5O7kPKoapBsikgx8Bpisqj5g\niYi8AFwGfKeL+0YCZwBXhhFzl+LczviGFE9K6zm/v5zKysVW2BgTBSUlJaSmpjJy5EhEpPsbOlFT\nU0NqamoPRta7TvT4VZUjR45QUlLCqFGjejRvUe18LqaIBOlkpHAH51VVw9kfZzqwVFWT2py7GZir\nqud3cd/twEdUdV6bc7uARJzmwDXAt1R1XQf3Xg1cDZCXlzfzhgf/wKPvNjIzR7m+OKV98pjg8/lI\nSYnN2CG244/l2KFv4k9PT2f06NHHVdAABAIB3O6Y7S6OifhVle3bt1NVVXXU+fnz569W1eJjzbe7\nwuEuen5lgBSgut25KqC74v5y4Iftzl2Ks5+O4AxmeFVEJqjqUT39qvo48DhAcXGxTpw4Ed5di8fj\nYd68ecf2KfrYokWLYjZ2iO34Yzl26Jv4N23aRFpaWvcJu3Gi1wy6Eyvxe71epk+f3qN5dreCwJ09\n+jSHj6M3YyN0XNPZDSJyOpAP/K3teVVd2ubwPhG5Aqep7Z9dBVDRUAHA/ganzyYQaGD16pmMGnU3\nubmfDu9TGGOMCVt382x2iMi0Hn7mVsAjImPbnJsGbOwkPcAVwHOhPp6uKB8sFNopb5wz9DnF43zD\nUG3E5UqgsXFvd7caY2LUPffcw6RJk5g6dSpFRUWsWLECv9/P9773PcaOHUtRURFFRUXcc889rfe4\n3W6KioqYNGkS06ZN44EHHiAY7LZr2nSgu2a0kTjDnHuMqtaKyHPAXSLyFZzRaBcAp3aUXkQSgc8B\nF7U7XwAMB1biFJrXAznA0vZ5tJfocTZNSwsVNh5POsXF7xzbBzLGnPCWL1/Oiy++yDvvvENCQgJl\nZWU0NTVx6623cvDgQdavX4/X66WmpoYHHvhgcndiYiJr164F4NChQ1xyySVUV1fzgx/8oK8+Sszq\ndrmaKLkOp2P/EPAMcK2qbhSRM0Skfe3lQqASWNjufCrwS6AC2Ad8HDhXVY909/CWQREa6o5SDdLY\nePCYP4wxJjJr1szjwIEnAQgGm1mzZh4HD/4BgECgjjVr5nHo0J8B8PurWLNmXusuuk1NZaxZM4+y\nMqe1PJz/dw8cOEBOTg4JCc5355ycHDIyMvjVr37Fww8/jNfrtHakpqZy5513dpjHoEGDePzxx/n5\nz39OVwOrTMfCKWx6/KeqquWqeqGqJqtqgar+MXT+LVVNaZf2GVUdoe3+dlV1o6pODeWRrapnqeqq\ncJ5fXu9MFTrQcACA2tr3WL/+XI4c+VdPfDxjzAnm7LPPZu/evYwbN47rrruOxYsXs23bNgoKCiLq\nsC8sLCQQCHDo0KEoRts/HcvmaZ2Jmc3TEuOcUddpcc44BREPHk82cXHZfRmWMQPG9OmLWt+7XHFH\nHbvdSUcdezzprcc1NTXEx+ccdT0hIb/b56WkpLB69WreeustFi5cyMUXX8z3vve9o9L89re/5aGH\nHuLIkSMsW7aM4cOHH8tHM50Ip7ApAhrDSBcz9cqkOKfPJjU0qTM5eQJFRREtQmCMiTFut5t58+Yx\nb948pkyZwmOPPcaePXtahyNfddVVXHXVVUyePJlAINBhHjt27MDtdjNo0KBejj72hVPYXKiqb0c9\nkl4UDLXIBa3d1ZgBYcuWLbhcLsaOdQbBrl27lvHjxzN9+nQWLFjAY489htfrJRAI0NTU1GEehw8f\n5pprrmHBggXHPTl1IBqQm6dVtOuzKSt7gR07vsuUKS+RmDiyDyMzxkSDz+fj+uuvp7KyEo/Hw5gx\nY3j88cdJT0/ntttuY/LkyaSmppKYmMgVV1zBkCFDAKivr6eoqIjm5mY8Hg+XXXYZ//u//9vHnyY2\nDcjCpqXPJj3UZ6Pqx+PJxOWynTqN6Y9mzpzJsmXLOrx2//33c//993d4rbPmNBO5gVnYhObZtCzE\nmZv7abKyzsHtTu7qNmOMMceou+Vq+moeTlQFQ4tTB9ssUm0FjTHGRE+/LEy6U1HvrI12sMGZDLZv\n3yOs+P/tnX90VdWVxz87IWkIIDEKyo8K+ItfCSGgAq2UpIrWNUXjaOsPtIItakXoUhbFJXWkHTuF\nstQWpIOdqWawNv2B/NCxOk6VCAxWIRAgiOKMBgQEA0jkVyAke/4498WXR0Ley8t7l5e3P2vd9d69\n79xzv/fk5e53zj5n73f6U19f66cswzCMdktSGpvM9IDPJpAWWunY8RJEknJU0TAMI+Yk5dM14LPp\n1MENnfXqNZlu3W6x6YyGYRgxIil7NnVe1NY6dTNN6uqOkZaW7ackwzCMdk1SGptAPpu9NXsB+OCD\nu1m3bpifkgzDiCGBVAE5OTmMGzeOgwcPtnxSGFRWVpKTk9MmdU2YMIF+/fo1pDqYN29em9TbFKWl\npc1OBY8VSWlsOqW7gMlHLQAAESdJREFU4bOs9Cz3mlVIVtZoPyUZhhFDAqkCKioqyM7OZsGCBX5L\napK5c+dSXl5OeXk5U6dODfu8SNcDmbGJExkdXJjxTqluokDPnvdwySXz/ZRkGElFQXEBxeXFANTW\n1VJQXMDvN7kUA0drj1JQXMCfKlyKgeqaagqKC1iy1aUY2Hd0HwXFBbz8gUsxsOdwZOlBRo0axa5d\nuwAXWeCqq65i2LBh5Obmsnz5csD1WAYOHMikSZMYPHgw11xzDceOHQOgrKyMvLw88vLyGhmtmpoa\nJk6cSG5uLvn5+axY4bKiFBcXU1RUxNixY8nJyeHpp5/mySefJD8/n5EjR3LgwIHT6i0pKSE3N5ec\nnBxmzJjRcLxz585MmzaNvLw83n77bcrKyhgzZgzDhw/n2muv5dNPXYSUefPmMWjQIIYMGcKtt95K\nZWUlCxcu5KmnnmLo0KGsWrUqovZrLUlpbAI+m5Nqq4MNI5moq6vjjTfe4PrrrwcgIyODpUuXsn79\nelasWMG0adMactV8+OGHTJ48mS1btpCVlcWLL74IwMSJE5k/fz4bN25sVPeCBQsQETZv3kxJSQl3\n3XUXNTU1AFRUVLBkyRJKS0uZOXMmmZmZbNiwgVGjRrFo0aKGOqZPn94wjLZ582Z2797NjBkzePPN\nNykvL2ft2rUsW7YMgCNHjjBixAg2btzIiBEjmDJlCosXL6asrIy7776bmTNnAi5CwoYNG9i0aRML\nFy6kb9++3HfffTz44IOUl5czenR8RnWScjbawRo3Xhvw2ZSXF5Ke3pNBg17wU5ZhJA2lE0ob3qel\npjXaz0zLbLTfNaNrw/6hQ4c4N/PcRp+f37nlFAOBGGe7du1i4MCBjB07FnCJFB955BFWrlxJSkoK\nu3btYu9e91wI+E/AhbuprKzk4MGDHDx4kG984xsA3Hnnnbz66qsArF69milTpgAwYMAA+vTpw7Zt\n2wAoLCxsyJvTtWtXxo0bB0Bubi6bNm1q0Dl37lxuvvnmhv3ly5dTUFBAt27dABg/fjwrV66kqKiI\n1NRUbrrpJsAFGq2oqGi4r7q6Onr06AHAkCFDGD9+PEVFRRQVFbXYVrEiKXs2ndIa+2yys68jK2uM\nn5IMw4ghAZ/N9u3bUdWG4a8XXniBqqoqysrKKC8v57zzzmvojQSyeoKbYHDy5MlWXz+4rpSUlIb9\nlJSUVtebkZFBamoq4Izm4MGDG/w9mzdv5vXXXwfglVdeYfLkyaxfv57LL788qvuIhqQ0Nl8J+Gw6\nOJ/NBRf8mJ497/FTkmEYcSAzM5N58+bxxBNPcPLkSaqrq+nevTtpaWmsWLGC7du3n/b8rKwssrKy\nWL16NeCMVYDRo0c37G/bto0dO3bQv3//qPReccUVvPXWW+zbt4+6ujpKSkoYM+bUH8b9+/enqqqK\nt99+G4Da2lq2bNlCfX09n3zyCYWFhcyZM4fq6moOHz5Mly5dOHToUFTaIiUpjU1dvfPVnKz3x8Ib\nhuEf+fn5DBkyhJKSEsaPH8+6devIzc1l0aJFDBgwoMXzn3vuOSZPnszQoUMJzlZ///33U19fT25u\nLrfccgvFxcWNejStoUePHsyePZvCwkLy8vIYPnw4N9xwwynl0tPTWbx4MTNmzCAvL4+hQ4eyZs0a\n6urquOOOOxomLUydOpWsrCzGjRvH0qVL4zpBQDTJEohddtlleu+vnuXn//kJF5y9h5Uzvs+aNb3o\n3v1WLr74Cb/lhU1paSkFBQV+y2g1iaw/kbWDP/q3bt3KwIEDo64nkFUzUUkU/U39vUSkTFUva22d\nSTlBoLO3zibbixrQo8ckunTJ91OSYRhGuyYpjU3AZxOIkdav3ywf1RiGYbR/ktJnU+v5amrra1FV\nkm0o0TD8wP7PEoNY/Z2S0th8UfMFAFXHP0P1BG+9lcL27b/wWZVhtF8yMjLYv3+/GZwzHFVl//79\nZGRktHndSTmM1jndpYPOTs8GUujT5zG6dr3SX1GG0Y7p3bs3O3fupKqqKqp6ampqYvIgjBeJoD8j\nI4PevXu3eb1JaWzSU9MB6JiaSUpKmvlsDCPGpKWl0a9fv6jrKS0tJT8/cSfzJLr+aPBlGE1EskVk\nqYgcEZHtInJ7M+VeFZHDQdsJEdkc9HlfEVkhIkdF5H0RuTqc65/UgM/mBKr11NUdRbW+Te7NMAzD\nOBW/fDYLgBPAecB44F9FZHBoIVW9TlU7BzZgDfCXoCIlwAbgHGAmsFhEurV08eqaagA+O/4ZNTWV\nrFrVib17n4/2ngzDMNolgSjZ0RB3YyMinYCbgEdV9bCqrgZeAu5s4by+wGhgkbd/KTAMeExVj6nq\ni8Bmr+7T0iXdLao6J/0cOnTI4sILZ9OlS6vXKhmGYbRbVJV777036nr88NlcCpxU1W1BxzYCLUXC\n/B6wSlUrvf3BwEeqGhzgZ6N3vBEicg8QCH52+PYRF30AsB3OTX+Ife7ww5Hdhf+cCwHtCUki609k\n7ZDY+hNZOyS2/qgCvflhbDoDX4QcqwZaiuHwPeDxkHqqm6inV+iJqvpb4Lehx0VkXTThF/wkkbVD\nYutPZO2Q2PoTWTsktn4RWRfN+X74bA4DZ4UcOwtoNgSpiFwJnA8sjqYewzAMwx/8MDbbgA4icknQ\nsTxgy2nOuQtYoqqHg45tAS4UkeAeUUv1GIZhGD4Qd2OjqkeAJcDPRKSTiHwduAFocjqYiHQEvgsU\nh9SzDSgHHhORDBG5ERgCvBiBnFOG1hKIRNYOia0/kbVDYutPZO2Q2Pqj0u5LigERyQaeBcYC+4GH\nVfUPIjIaeNWb5hwoexswG+irIWK9GWrFwAhgBzBZVf8Wj3swDMMwwifp8tkYhmEY8ScpA3EahmEY\n8cWMjWEYhhFz2rWxiSAGm4jIHBHZ721zRETirTdEU7jaC734cNUiUhlnmc0Sgf7pIlIhIodE5GMR\nmR5vrU1oClf7gyLykYh8ISK7ReQpEfE9uG24+oPKp4vIVhHZGS+Np9ESbtvPEpHakNiJF8ZbbxO6\nwm57ERkmIis97XtF5Efx1NqEnjaJWdkc7drYEGYMNlx0gSLc1OkhwDgg+vgM0RGu9iO4yRa+P6RD\nCFe/4Bbsng18C3hARG6Nm8qmCVf7S8AwVT0LyMF9f6bGTWXzhKs/wHQgutj/bUck2v8UHDtRVT+K\nm8rmCUu/iJwLvAY8g4vteDHwehx1NkVbxaxsmkCmyva2AZ28hrs06NjzwOwmyq4B7gna/z7w90TQ\nHvT51UCl3+3eWv1B5eYB8xNNO+6B8TfgN4nU9kA/YCtwHbAzUbQDs4Df+6k3Sv3/Ajzvt+bWfm+C\nyvQF6nCzhU97jfbcs2kuBltTv5IGe5+1VC5eRKL9TKRV+r2hy9H4uzA3Iu0icruIfIGLd5WH+6Xq\nJ5G2/XzgEeBYrIWFQaTax4nIARHZIiI/jL28FolE/0jggIisEZHPRORlEbkgLiqbprXPnNCYlc3S\nno1NJDHYQuOsVQOdffTbtDZ+3JlCa/XPwn0nn4uBpnCJSLuq/kHdMNqlwEJgb2zltUjY+r2F0Kmq\nujQewsIgkrb/MzAQ6AZMAv7JW5PnJ5Ho742LjPIj4ALgY1zKFL+IJmZlcTgXaM/GJpLYaaFlzwIO\nq9dP9IFEj/vWmvh3D+C+uP+gqsdjqK0lWtX2qvohrkf2mxjpCpew9ItL9fFLzgwfU4Cw215V31PV\n3apap6prgF8DN8dB4+mI5LtzDFiqqmtVtQb4KfA1EekaY43N0VYxK5ulPRubSGKwbfE+a6lcvGhN\n/LgziYj0i8jduBwPV6mq3zOiomn7DsBFMVEVPuHqvwQ33r5KRPbgQkj1EJE94iJz+EE0ba+4ySZ+\nEon+TTjNAfxeXd9WMSubx2/HVIydXn/EdU07AV/HdQsHN1HuPpyTtBfQ02vg+xJEewqQgXPwbvfe\npydQ248H9gAD/dbcCu0/ALp77wd535snE0E/zjCeH7T9I7Dbe596Jmv3yt2Am8EowBXALuCuRGh7\nr9w3gc+BoUAa8BTO93HGa/fKdvQ+/2bY9fv9x4lx42UDy3DTg3cAt3vHR+OGyQLlBDekcMDbfokX\nyicBtBfgfhUFb6UJ1PYfA7W4bnxgW5gg2p/D+WiOAJXAXCAjUdo+5JwCfJ6NFmHbl+DiKh4G3gem\n+q090rYHfogzkp8DLwNfTSDtt+F+3Ib9nLTYaIZhGEbMac8+G8MwDOMMwYyNYRiGEXPM2BiGYRgx\nx4yNYRiGEXPM2BiGYRgxx4yNYRiGEXPM2BjtGhGZICLazHa1V6ZvyPETIrLNy09zdlBds0LKHReR\n97ycPGH/L4nIQyKyKTj2nnfsZRH51Kt7Vps2xJfXGSUi73g5S1REhjZxXye9fCa/E5FeIef/SkT+\nGgttRvvG90RPhhEnvgOEhsJ5L2T/F7gcNV/BraD+CZAvIoXaeEHalbiw6tnABNwi4HrgiZZEiEgW\nMBO4N6TOSbhAiMtwES1ixe9wcbnGAUdxYUoCBO4rDRcR4afAcBEZpqr1Xpk5wEdem6yIoU6jnWHG\nxkgWylX1f1so85Gq/t17/5aIpOEiUecD64PKvaOqJwFE5DVcwr1JhGFscLmSTgChkZYHq2q9uEyf\nMTE2Xu+rP/BzVX0z6HjgbcN94WKm1QH/5p2zFUBVPxWRl3EJ18zYGGFjw2iG0TxrvdeLmyvg/eLf\niAsTHw4/AP6sqnVN1NNqRCRTXDrzj71hwI9FZGZgeE9EJuB6LSnAo95wWWUL1QZCzqeFHP8jcK2I\nfDUazUZyYcbGSBZSRaRD0JYaxjn9vNeDLZTrC/xfS5WJSB9gALAqjGuHjdcb+i+cIfs1LijrvwOP\n4uK1AbyCGyYDN5Q2CrgxpKpAG3UUkeG4pGpbgIqQcqtwz46xbXkfRvvGhtGMZOH9kP3/4cuHb4AU\n78Gdzpc+m0851TikekNPZ+Me8MMJL5fKSO9142lLRc5tuHsZo6orvWNveBofE5E5qvqZiHzufbYz\naLgwmJqQ/feBb4f2ulS1SkR24u7n2Ta7C6NdY8bGSBZupPEEgaaSQj1D47TOq4HJqhqaMjn0ofxj\nVV0Whoae3mtVGGVPwRsSCx6NqPcMwbdwEXjXeMYywOvA4zij8FIYlxjJl0NtfYAZwOsi8jVVDc1A\nWsWX92MYLWLGxkgWKsKYIPA4sBw4DuxQ1epmyo3EzT7rhRuqmi0ia1W1tIX6M7zX1mYifRaXsCrA\nf+Bmw3XHGYfaZs47J8z6y4ImCLwrIitxPbuHcIYnmGO4nCaGERZmbAzjS7ar6rowygUeymtFZDVu\nuGm+iOS14Ojf772ejXtYR8os4Omg/X1B9X4MfLeZ8ypbcS1Uda+I7MPNtgslG5dt0jDCwoyNYUSB\nqu4TkZ/hHPM3AX85TfGA3+hCXFbMSK9VSdOG4zXv2odVNdQ31WpEpAdwLiHDft7kigs4/b0aRiPM\n2BhG9DyDW3fyExFZHLJYM5h3cUNoV+D8QQ2IyGW4WW0Bn8wgEQlMOvirqh49zfVfACbiJgU8gZuA\nkA5cBFwPFLVwfoAR3tqagM9mOs6HszCkXA6QCazEMMLEjI1hRImqHheRf8YZnSJOXbAZKFcjIstx\nq/efDPn4ARr7Y77jbeCmYFee5vq1InIt8DBwj1f+CG469iu4RaThEDCACuwByoD7VPXdkHLf9j4v\nDbNew7C00IYRT0SkAHgT6KuqO3yW0ypE5D3gRVV91G8tRuJgxsYw4oyI/Dfwgao+4LeWSBGRG3Cz\n4i5S1ZYWuxpGAxZBwDDizxRgZ3DU5wSiI3CHGRojUqxnYxiGYcQc69kYhmEYMceMjWEYhhFzzNgY\nhmEYMceMjWEYhhFzzNgYhmEYMef/AevXyY+kSCF5AAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-CXnpwbUB6hB",
        "colab_type": "code",
        "outputId": "eb8d4459-a808-4faf-9954-f05540b7fc35",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 290
        }
      },
      "source": [
        "plt.plot(prec_forest,prec_forest*rec_forest,\"g:\",label=\"RandomForest\")\n",
        "plt.plot(precisions,precisions*recalls,\"y:\",label=\"SGD\")\n",
        "plt.plot(prec_mlp,rec_mlp*prec_mlp,\"blue\",label=\"MLP\")\n",
        "plt.grid(True) \n",
        "plt.xlabel(\"pureza\")\n",
        "plt.ylabel(\"eficiência * pureza\")\n",
        "plt.legend(loc=\"upper left\")\n",
        "plt.show()"
      ],
      "execution_count": 62,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYkAAAERCAYAAACO6FuTAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "oIoibpJKOsVM",
        "colab_type": "text"
      },
      "source": [
        "Para achar o ponto com o melhor valor de threshold:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VFcWRMHU8wzw",
        "colab_type": "code",
        "outputId": "2aca0017-4a68-4d8e-e33d-79083429ae61",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 327
        }
      },
      "source": [
        "idxmaxeffpur = np.argmax(prec_mlp*rec_mlp)\n",
        "#effpurmax = np.max(prec_mlp*rec_mlp)\n",
        "#print(effpurmax,idxmaxeffpur)\n",
        "print(\"Melhor Eficiencia: \",rec_mlp[idxmaxeffpur],\" Melhor precisão: \",prec_mlp[idxmaxeffpur])\n",
        "best_thresh_mlp=thresh_mlp[np.argmax(prec_mlp>=prec_mlp[idxmaxeffpur])]\n",
        "print(\"Melhor valor do threshold: \",best_thresh_mlp)\n",
        "y_pred_mlp_best = (y_pred_mlp>= best_thresh_mlp)\n",
        "f1_score(y_train_5,y_pred_mlp_best)\n",
        "plt.hist(y_pred_mlp[y_train_5] ,bins=100,range=(0,1),histtype='step',label=\"sinal\")\n",
        "plt.hist(y_pred_mlp[~y_train_5],bins=100,range=(0,1),histtype='step',label=\"ruido\")\n",
        "plt.plot([best_thresh_mlp,best_thresh_mlp],[0.01,1e5],\"r:\",label=\"melhor threshold\")\n",
        "plt.axis([-0.05,1.05,0.01,1e5])\n",
        "plt.yscale(\"log\")\n",
        "plt.xlabel(\"Discriminador\")\n",
        "plt.legend(loc=(0.58,0.75),frameon=False)\n",
        "plt.show()"
      ],
      "execution_count": 114,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Melhor Eficiencia:  0.9681919642857143  Melhor precisão:  0.9897318881916715\n",
            "Melhor valor do threshold:  0.5188121\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYEAAAEUCAYAAADN8orUAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    }
  ]
}