bestRFmixTPC4ICA_RAWCODE.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Copy of bestRFmixTPC4ICA.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.7.6"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "rBtv-qQV-ZUS"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "-02tL1nQFQrk",
        "outputId": "5e4f5df6-d6a8-4623-b44c-5f295fc34ce0"
      },
      "source": [
        "%%html\n",
        "<iframe src=\"https://icarus.fnal.gov/\" width=\"1200\" height=\"1000\"></iframe>"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "<iframe src=\"https://icarus.fnal.gov/\" width=\"1200\" height=\"1000\"></iframe>\n"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 39
        },
        "id": "ZCUpKGHhAR7s",
        "outputId": "a6e1b1e4-5cb3-4275-cee5-f9d7b03d3018"
      },
      "source": [
        "%%html\n",
        "<marquee style='width: 55%; color: purple;'><b>RANDOM FOREST</b></marquee>"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "<marquee style='width: 55%; color: purple;'><b>RANDOM FOREST</b></marquee>\n"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yx0QUuR6-aFZ"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Z-TxKqJ0SAbW"
      },
      "source": [
        " #only in colab:\n",
        "\n",
        "from google.colab import drive\n",
        "drive.mount('/content/drive/')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9WieqQ8F5o89"
      },
      "source": [
        "#sigdir = /content/drive/MyDrive/nu\n",
        "#backdir = /content/drive/MyDrive/cosmic"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "q6iVBXfBazUN"
      },
      "source": [
        "sigdir = '/content/drive/MyDrive/nu'\n",
        "backdir = '/content/drive/MyDrive/cosmic'"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_Bl3rvzY5ddp"
      },
      "source": [
        "print('Insert path directoty of \\033[91m\\033[1mSignal files\\033[0m')\n",
        "#sigdir = input('')\n",
        "print('Insert path directoty of \\033[91m\\033[1mBackground files\\033[0m')\n",
        "#backdir = input('')\n",
        "#print('\\033[1m\\033[94mtpc selection:\\033[0m')\n",
        "#tpc = int(input())\n",
        "#print('\\033[1m\\033[94mplane selection:\\033[0m')\n",
        "#plane = int(input())"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9Jo2AihNJ6pE"
      },
      "source": [
        "# Extracting  the raw data\n",
        "The raw data(in contex of ICT) is provided as .root file. Hence, to extract this data type to python enviroment, we used Uproot library. Uproot ver. 3.9 is very efficient for our purposes as it could be convert to pandas data frame."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "n8qGVdAO-2MZ"
      },
      "source": [
        "![rootuproot.png](data:image/png;base64,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)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XZOvF89h-ta0"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "e3fO4apqSEAD"
      },
      "source": [
        "!pip install awkward\n",
        "!pip install uproot==3.9"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "eTfzY6ORdsLQ"
      },
      "source": [
        "## importing required libraries"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5HwZUXZ1WqeV"
      },
      "source": [
        "import scipy as sp\n",
        "import sklearn\n",
        "from sklearn.datasets import fetch_openml\n",
        "from sklearn.neural_network import MLPClassifier\n",
        "from sklearn.decomposition import PCA\n",
        "import sklearn.metrics as skm"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "u9qYa7VmW-1b"
      },
      "source": [
        "import gzip\n",
        "from sklearn.decomposition import IncrementalPCA\n",
        "import seaborn as sns"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UAZwnkCzSEDU"
      },
      "source": [
        "import pkg_resources\n",
        "import uproot\n",
        "import os\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import math\n",
        "import matplotlib.pyplot as plt\n",
        "%matplotlib inline\n",
        "import plotly.graph_objects as go\n",
        "import re\n",
        "import awkward\n",
        " \n",
        "import itertools\n",
        "from mpl_toolkits.mplot3d import Axes3D\n",
        "from scipy import linalg as la\n",
        " \n",
        "from sklearn.neural_network import MLPClassifier\n",
        "from sklearn.model_selection import GridSearchCV\n",
        "from sklearn.model_selection import ParameterGrid\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.preprocessing import MinMaxScaler\n",
        "from sklearn.metrics import accuracy_score\n",
        "from sklearn.ensemble import RandomForestClassifier\n",
        "from sklearn.metrics import roc_curve\n",
        "from sklearn.metrics import roc_auc_score\n",
        " \n",
        "from tqdm import tqdm\n",
        "# import alive_progress\n",
        "import time\n",
        "seed = 666\n",
        "np.random.seed(seed)\n",
        "from sklearn.metrics import confusion_matrix\n",
        "from sklearn.metrics import roc_curve\n",
        "from sklearn.metrics import roc_auc_score\n",
        "from sklearn.metrics import accuracy_score, confusion_matrix, classification_report\n",
        "from sklearn.model_selection import RepeatedStratifiedKFold\n",
        "from sklearn.model_selection import cross_val_score\n",
        "from sklearn.metrics import confusion_matrix\n",
        "import seaborn as sns\n",
        "import glob"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hFKcJybw0Ay1"
      },
      "source": [
        "from sklearn.model_selection import train_test_split"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Nm5P6wfqYCXF"
      },
      "source": [
        "from sklearn.svm import SVC\n",
        "\n",
        "from sklearn.model_selection import GridSearchCV"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vlKjvgGLodM9"
      },
      "source": [
        "from math import sqrt"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ayO7tgbuXQKs"
      },
      "source": [
        "np.random.seed(777)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rZQ06vT0eGx4"
      },
      "source": [
        "### print_full(x) \n",
        "it put here and its job is to print whole a data frame"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "eQ54-_mbSwTL"
      },
      "source": [
        "def print_full(x):\n",
        "    pd.set_option('display.max_rows', len(x))\n",
        "    print(x)\n",
        "    pd.reset_option('display.max_rows')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TjUyR6peeuFb"
      },
      "source": [
        "## Features of input data\n",
        "Based on the priore knoledge, and also with looking at the correlation matrix of the features of raw data, these features are selected."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LBZSQ9ipSEGk"
      },
      "source": [
        "colmns = ['hit_tpc', 'hit_plane', 'no_hits', 'hit_wire', 'hit_peakT']"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4aJmou4M5JLc"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XVIANmXQfhnN"
      },
      "source": [
        "## Printing out the data from files\n",
        "Changing the working directory to nues files path directory. \"sigdir\" represented nue files which inserted before."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dj24JIDNS0Si"
      },
      "source": [
        "os.chdir(sigdir)\n",
        "os.getcwd()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PcESk1q1g-Ws"
      },
      "source": [
        "Calling all nue files where placed in a folder with glob.glob."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vqEYJ4kqS2_d"
      },
      "source": [
        "allFiles_mu = os.listdir()\n",
        "allFiles_mu = glob.glob('./*.root')\n",
        "\n",
        "allFiles_mu.sort()\n",
        "allFiles_mu"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9S_e81XE_UV7"
      },
      "source": [
        "# nue_400 contains some deferent tree files!!!\n",
        "#allFiles_mu.remove('./nue_400.root')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "YdUy4YJbhgSA"
      },
      "source": [
        "## ROOT files\n",
        "Root files use TTree to store ascii files. nue files are simolation of neutrino events(shortbase line) with keword 'icaana' and for branch is 'anatree;1'. By using Uproot 3.9, the files exxtracted as build-in Awkward array and then converted to pandas data frame for each nue files in the file."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TDKy2hjLSEJv"
      },
      "source": [
        "df = []\n",
        "for file in allFiles_mu:\n",
        "    tt = uproot.open(file)['icaana']['anatree;1']\n",
        "    pf = tt.pandas.df((colmns))\n",
        "    df.append(pf)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "IHlltlvBmDpa"
      },
      "source": [
        "### Data Integration\n",
        "The nue files(it is true about cosmic files) sould be combined. Each nue file contains from 100 entrys counted from 0 to 99. Therefore, it is required another index to identifiable entrys. It is done by putting the file names."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "i9edVb9YZSIo"
      },
      "source": [
        "df_con = pd.concat([df[i] for i in range(19)],ignore_index=False, keys=allFiles_mu)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GsabwM75ZbD-"
      },
      "source": [
        "df_con"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PEPFpHrdsVPj"
      },
      "source": [
        "### Labeling\n",
        "From Machine learning aspect, our problem is the binary classification. So we have to determine the target, here is nue files. For unsupervised learning, tagging is not required."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "czC0XfBFTCMY"
      },
      "source": [
        "tag = 'ev'\n",
        "labl = 1\n",
        "df_con['label'] = labl\n",
        "df_con['tagg'] = tag\n",
        "\n",
        "df_con"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "K5EU6CGDtfil"
      },
      "source": [
        "## Data Cleaning\n",
        "we worked with Pandas. Hence, for cleaning and sanitaizing data from NA and NAN, .dropna() has been exploited."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4DXpltwlTFhr"
      },
      "source": [
        "df_con = df_con.dropna()\n",
        "print(df_con.info())"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LUZ1lM7kuUmD"
      },
      "source": [
        "# Data reduction\n",
        "Based on geometry of LAr TPC, the first induction of TPC has higher resolution i.e it has lower systematic uncertainity/noise in comparison with other area of TPC. In this notebook, the selected input data is provided by data beloning to collection and tpc 0 (hit_plane = 2, hit_tpc = 0)  and induction 2 and tpc 1 (hit_plane = 1, hit_tpc = 1)."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FPaCBr7ITJKt"
      },
      "source": [
        "df_intpc0 = df_con.loc[df_con['hit_tpc'] == 0]\n",
        "df_intpc1 = df_con.loc[df_con['hit_tpc'] == 1]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-I92rTd1v_BH"
      },
      "source": [
        "To combine these two area, it should be minded the mirroring effect in result of location of the anode and the cathod coils"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6M-AnktCTMxJ"
      },
      "source": [
        "rep0 =df_intpc0.hit_peakT - 865\n",
        "rep1 =-df_intpc1.hit_peakT + 5505"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dHTGv6ETSEM3"
      },
      "source": [
        "df_intpc0.iloc[:, 4] = rep0\n",
        "df_intpc1.iloc[:, 4] = rep1"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "iqwrHpDFSEQD"
      },
      "source": [
        "df_in0_2 = df_intpc0.loc[df_intpc0['hit_plane'] == 2]\n",
        "df_in1_1 = df_intpc1.loc[df_intpc1['hit_plane'] == 1]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sX3xVcxfSETr"
      },
      "source": [
        "df_in0_2"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gvekcbTWYk7f"
      },
      "source": [
        "df_in1_1"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8cXmQWd2SEYY"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MvhQDz57w6mu"
      },
      "source": [
        "### Cosmic Files\n",
        "The procedure is the same as it has been furbished for the nue files."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Bvu16YAeSEcT"
      },
      "source": [
        "os.chdir(backdir)\n",
        "os.getcwd()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "eeBBFk16SEfS"
      },
      "source": [
        "FLocosmic = os.listdir()\n",
        "FLocosmic = glob.glob('./*.root')\n",
        "\n",
        "FLocosmic.sort()\n",
        "FLocosmic"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VZYjIloBT356"
      },
      "source": [
        "dfc = []\n",
        "for file in FLocosmic:\n",
        "    ctt = uproot.open(file)['icaana']['anatree;1']\n",
        "    cpf = ctt.pandas.df((colmns))\n",
        "    dfc.append(cpf)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "srarWn-K_fbn"
      },
      "source": [
        "dfc_con = pd.concat([dfc[i] for i in range(37)],ignore_index=False, keys=FLocosmic)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yOeZdzmUVTCK"
      },
      "source": [
        "lbl = 0\n",
        "dfc_con['label'] = lbl"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1PSovj8HSElZ"
      },
      "source": [
        "tagi = 'cos'\n",
        "dfc_con['tagg'] = tagi"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "l03WOXVnVVRd"
      },
      "source": [
        "dfc_intpc0 = dfc_con.loc[dfc_con['hit_tpc'] == 0]\n",
        "dfc_intpc1 = dfc_con.loc[dfc_con['hit_tpc'] == 1]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Fi3ymV0VVX7S"
      },
      "source": [
        "repc0 =dfc_intpc0.hit_peakT - 865\n",
        "repc1 =-dfc_intpc1.hit_peakT + 5505"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lhdm0rjMVZ-v"
      },
      "source": [
        "dfc_intpc0.iloc[:, 4] = repc0\n",
        "dfc_intpc1.iloc[:, 4] = repc1"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zP1ZKgM0Vc7q"
      },
      "source": [
        "dfc_in0_2 = dfc_intpc0.loc[dfc_intpc0['hit_plane'] == 2]\n",
        "dfc_in1_1 = dfc_intpc1.loc[dfc_intpc1['hit_plane'] == 1]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rVlslTgDVoF1"
      },
      "source": [
        "dfc_in1_1"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GPtV9MUCVgpT"
      },
      "source": [
        "data = pd.concat([df_in0_2, df_in1_1, dfc_in0_2, dfc_in1_1])"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mehHlkQE2oqW"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0BBMwQPn2o-e"
      },
      "source": [
        "##Scatter plots of data"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7PyXkO4I2vfr"
      },
      "source": [
        "###Scatter plots of nue - mix tpc"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UcdoxtHl210V"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KeIBKnM5zIqy"
      },
      "source": [
        "for i in range(19):\n",
        "\n",
        "  for j in range(100):\n",
        "    plt.figure(figsize=(11,16))\n",
        "    plt.title(allFiles_mu[i]+', entry: '+str(j), fontsize=18)\n",
        "    plt.xlabel('hit_wire', fontsize=14)\n",
        "    plt.ylabel('hit_peaktime', fontsize=14)\n",
        "    pdf = df[i].iloc[df[i].index.get_level_values('entry') == j]\n",
        "    p1 = pdf['hit_plane'] ==1\n",
        "    ii = pdf['hit_plane'] ==2\n",
        "    zo = pdf['hit_tpc'] ==0\n",
        "    un = pdf['hit_tpc'] ==1\n",
        "    #plt.scatter(pdf['hit_wire'],pdf['hit_peakT'], color = \"white\", marker = \".\")\n",
        "    plt.scatter(pdf['hit_wire'].loc[ii&zo],pdf['hit_peakT'].loc[ii&zo]-865, color = 'black')\n",
        "    #plt.scatter(pdf['hit_wire'].loc[p1&zo],pdf['hit_peakT'].loc[p1&zo], color = 'red')\n",
        "    plt.scatter(pdf['hit_wire'].loc[p1&un],pdf['hit_peakT'].loc[p1&un]*(-1)+5505, color = 'red')\n",
        "    #lt.scatter(pdf['hit_wire'].loc[p1&un],pdf['hit_peakT'].loc[p1&un], color = 'green')\n",
        "    ll = ['plane=2, tpc=0','plane=1, tpc=1']\n",
        "    plt.legend(ll, loc=\"upper left\", title=\"Classes\")\n",
        "    plt.show()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tlFlr1Go3EjA"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-sYHaCNf6yk2"
      },
      "source": [
        "plt.figure(figsize=(11,16))\n",
        "plt.title('nue_all', fontsize=18)\n",
        "plt.xlabel('hit_wire', fontsize=14)\n",
        "plt.ylabel('hit_peaktime', fontsize=14)\n",
        "for i in range(19):\n",
        "  for j in range(250):\n",
        "    pdf = df[i].iloc[df[i].index.get_level_values('entry') == j]\n",
        "    p1 = pdf['hit_plane'] ==1\n",
        "    ii = pdf['hit_plane'] ==2\n",
        "    zo = pdf['hit_tpc'] ==0\n",
        "    un = pdf['hit_tpc'] ==1\n",
        "    plt.scatter(pdf['hit_wire'],pdf['hit_peakT'], color = \"black\", marker = \".\")\n",
        "    \n",
        "    \n",
        "    plt.scatter(pdf['hit_wire'].loc[ii&zo],pdf['hit_peakT'].loc[ii&zo]-865, color = 'red')\n",
        "    plt.scatter(pdf['hit_wire'].loc[p1&un],-pdf['hit_peakT'].loc[p1&un]+5505, color = 'blue')\n",
        "\n",
        "    ll = ['glob', 'plane=2, tpc=0','plane=1, tpc=1']\n",
        "    plt.legend(ll, loc=\"lower right\", title=\"Classes\")\n",
        "    plt.grid('on')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Y-jIqIXw3Ft8"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LrvO0bZF3NXh"
      },
      "source": [
        "for i in range(37):\n",
        "\n",
        "  for j in range(250):\n",
        "    plt.figure(figsize=(11,16))\n",
        "    plt.title(FLocosmic[i]+', entry: '+str(j), fontsize=18)\n",
        "    plt.xlabel('hit_wire', fontsize=14)\n",
        "    plt.ylabel('hit_peaktime', fontsize=14)\n",
        "    pdf = df[i].iloc[df[i].index.get_level_values('entry') == j]\n",
        "    p1 = pdf['hit_plane'] ==1\n",
        "    ii = pdf['hit_plane'] ==2\n",
        "    zo = pdf['hit_tpc'] ==0\n",
        "    un = pdf['hit_tpc'] ==1\n",
        "    #plt.scatter(pdf['hit_wire'],pdf['hit_peakT'], color = \"white\", marker = \".\")\n",
        "    plt.scatter(pdf['hit_wire'].loc[ii&zo],pdf['hit_peakT'].loc[ii&zo]-865, color = 'green')\n",
        "    #plt.scatter(pdf['hit_wire'].loc[p1&zo],pdf['hit_peakT'].loc[p1&zo], color = 'red')\n",
        "    plt.scatter(pdf['hit_wire'].loc[p1&un],pdf['hit_peakT'].loc[p1&un]*(-1)+5505, color = 'blue')\n",
        "    #lt.scatter(pdf['hit_wire'].loc[p1&un],pdf['hit_peakT'].loc[p1&un], color = 'green')\n",
        "    ll = ['plane=2, tpc=0','plane=1, tpc=1']\n",
        "    plt.legend(ll, loc=\"upper left\", title=\"Classes\")\n",
        "    plt.show()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sMgSN-tv57hs"
      },
      "source": [
        "plt.figure(figsize=(11,16))\n",
        "plt.title('nues', fontsize=18)\n",
        "plt.xlabel('hit_wire', fontsize=14)\n",
        "plt.ylabel('hit_peaktime', fontsize=14)\n",
        "for i in range(37):\n",
        "  for j in range(250):\n",
        "    pdf = df[i].iloc[df[i].index.get_level_values('entry') == j]\n",
        "    p1 = pdf['hit_plane'] ==1\n",
        "    ii = pdf['hit_plane'] ==2\n",
        "    zo = pdf['hit_tpc'] ==0\n",
        "    un = pdf['hit_tpc'] ==1\n",
        "    #plt.scatter(pdf['hit_wire'],pdf['hit_peakT'], color = \"black\", marker = \".\")\n",
        "    \n",
        "    #plt.scatter(pdf['hit_wire'].loc[p1&zo],pdf['hit_peakT'].loc[p1&zo]-865, color = 'red')\n",
        "    #plt.scatter(pdf['hit_wire'].loc[p1&un],-pdf['hit_peakT'].loc[p1&un]+5505, color = 'green')\n",
        "    plt.scatter(pdf['hit_wire'].loc[ii&zo],pdf['hit_peakT'].loc[ii&zo]-865, color = 'blue')\n",
        "    plt.scatter(pdf['hit_wire'].loc[p1&un],-pdf['hit_peakT'].loc[p1&un]+5505, color = 'green')\n",
        "    ll = ['plane=2, tpc=0','plane=1, tpc=1']\n",
        "    plt.legend(ll, loc=\"lower right\", title=\"Classes\")\n",
        "    plt.grid('on')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cZ4cN2aAVpAl"
      },
      "source": [
        "datal = data.drop(['tagg', 'hit_tpc',\t'hit_plane'], axis=1)\n",
        "datal"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "94Yg3GEoxoqD"
      },
      "source": [
        "# Traning the Data"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Ri8-xcWnyR2D"
      },
      "source": [
        "## Data segmentation\n",
        "For traning the data, usually the data split out to two subsets. It is done by train_test_split of Scikit-learn package."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jJE9QjP_F_4g"
      },
      "source": [
        "X = datal.drop(labels=['label'], axis = 1)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zOzQfms3GYGw"
      },
      "source": [
        "Y = datal['label']"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rgOLNymRGYMk"
      },
      "source": [
        "X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size = 0.2, random_state = 76)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZMXyoTVeGYSw"
      },
      "source": [
        "X_train"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rB742muD0QWr"
      },
      "source": [
        "### Receiver Operating Characteristic - ROC Curve Analysis\n",
        "To evaluate the model(Random Forest), it is important to find true positive, false positive, true negative and false negative. By extracting them, it is easy to talk about the model as it means that concepts like selectivity, sensitivity, precision, miss rate and prevalence threshold and false omission rate are compromised."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WJPZrLT_GYY6"
      },
      "source": [
        "# funciton for ROC analysis\n",
        "class ROC_Analysis():\n",
        "    \n",
        "    def __init__(self, name, true_labels, predictions, fpr_th=[0.01, 0.02, 0.05, 0.1, 0.2], model=None):\n",
        "        \n",
        "        self.name = name\n",
        "        self.true_labels = true_labels\n",
        "        self.predictions = predictions\n",
        "        self.fpr_th = fpr_th\n",
        "        \n",
        "        self.fpr = None\n",
        "        self.tpr = None\n",
        "        self.thresholds = None\n",
        "        self.AUC_score = None\n",
        "        \n",
        "        self.model = model\n",
        "        \n",
        "        \n",
        "    def Compute_ROC_Curve(self):\n",
        "        \n",
        "        self.fpr, self.tpr, self.thresholds = roc_curve(self.true_labels, self.predictions)\n",
        "        self.AUC_score = roc_auc_score(self.true_labels, self.predictions)\n",
        "        \n",
        "        \n",
        "    def Find_Closest_Parameters(self):\n",
        "        \n",
        "        s = len(self.fpr_th)\n",
        "        closest_fpr, closest_tpr, indexes = np.zeros(s), np.zeros(s), np.zeros(s, dtype=int)\n",
        "\n",
        "        for ii in range(s):       \n",
        "            indexes[ii] = np.searchsorted(self.fpr, self.fpr_th[ii])\n",
        "        \n",
        "        closest_fpr = self.fpr[indexes]\n",
        "        closest_tpr = self.tpr[indexes]\n",
        "        closest_th  = self.thresholds[indexes]\n",
        "        \n",
        "        return (closest_fpr, closest_tpr, closest_th)\n",
        "    \n",
        "        \n",
        "    def Plot_ROC_Curve(self, axes=None):\n",
        "        \n",
        "        if self.AUC_score == None: self.Compute_ROC_Curve()\n",
        "        \n",
        "        if axes == None: fig, axes = plt.subplots(figsize=(7,6))\n",
        "        axes.plot(self.fpr, self.tpr, color=\"blue\", label=self.name+\"_ROC\")\n",
        "        axes.plot([0, 1], [0, 1], color=\"darkred\", linestyle='--', label=\"Random Guess\")\n",
        "        \n",
        "        cl_fpr, cl_tpr, cl_th = self.Find_Closest_Parameters()\n",
        "        axes.scatter(cl_fpr, cl_tpr, color=\"violet\", label=\"Relevant pts, with thr. values\")\n",
        "        for ii in range(len(self.fpr_th)): \n",
        "            plt.annotate('%.3f' % cl_th[ii], (cl_fpr[ii], cl_tpr[ii]), \n",
        "                         xytext=(cl_fpr[ii]+0.03,cl_tpr[ii]-0.02),fontsize=11)\n",
        "        \n",
        "        txt_roc_auc = \"roc_auc = %.3f\" % self.AUC_score\n",
        "        props = dict(boxstyle=\"round\", facecolor=\"gold\", alpha=0.5)\n",
        "        axes.text(0.6, 0.4, txt_roc_auc, transform=axes.transAxes, fontsize=12, verticalalignment='top', bbox=props)\n",
        "\n",
        "        axes.set_xlabel(\"False Positive\")\n",
        "        axes.set_ylabel(\"True Positive\")\n",
        "        axes.set_title(\"ROC Curve of the \" + self.name)\n",
        "        axes.grid()\n",
        "        axes.legend()\n",
        "        \n",
        "    \n",
        "    def Display_Results(self):\n",
        "        \n",
        "        if self.AUC_score == None: self.Compute_ROC_Curve()        \n",
        "        cl_fpr, cl_tpr, cl_th = self.Find_Closest_Parameters()\n",
        "        \n",
        "        inf_df = pd.DataFrame(np.vstack((np.array(self.fpr_th), cl_fpr, cl_tpr, cl_th)).round(4))\n",
        "        inf_df.index = ['values', 'False Positive','True Positive','threshold']\n",
        "        inf_df.columns = ['']*5\n",
        "        \n",
        "        return inf_df\n",
        "            "
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ScqL1Cg572XS"
      },
      "source": [
        "## Searching for the best RF\n",
        "The Grid search here has been reconciled. The most accordant RF would be a forest contains trees with min_samples_leaf = 1 and min_sample_split = 10 and the population of the forest would be 42 trees. It is necessary to mention that machine has been taghth in subentry level i.e the data has been comfronted disregarding the physical events."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jDYzoYtVGYev"
      },
      "source": [
        "def RFF(bootstrap,max_features,min_samples_leaf,min_samples_split,n_estimators):\n",
        "       \n",
        "    return RandomForestClassifier( bootstrap=bootstrap, max_features=max_features, \n",
        "               min_samples_leaf=min_samples_leaf, min_samples_split=min_samples_split, \n",
        "               n_estimators=n_estimators)\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rJPQl4q_IGIQ"
      },
      "source": [
        "pararafo = {'bootstrap': [True, False],'max_features': ['auto', 'sqrt'],\n",
        "             'min_samples_leaf': [1],  'min_samples_split': [10 ],\n",
        "             'n_estimators': [37,47]}"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KURHePPLvtIW"
      },
      "source": [
        "pararafo"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FS3PmsxAn5BN"
      },
      "source": [
        "bulding_forest_start=time.time()\n",
        "best_RF = {}\n",
        " \n",
        "counter = 1\n",
        "for params in tqdm(list(ParameterGrid(pararafo))):\n",
        "    rf = RFF(**params)\n",
        "    rf.fit(X_train, Y_train)\n",
        "    pred = rf.predict_proba(X_test)[:, 1]\n",
        "    best_RF['RF_'+str(counter)] = ROC_Analysis(name='RF_'+str(counter),  true_labels=Y_test, predictions=pred, model=rf)\n",
        "    counter+=1\n",
        "bulding_forest_end=time.time()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0WpJrGXyjOSR"
      },
      "source": [
        "t1 = bulding_forest_end - bulding_forest_start\n",
        "print('wall time of building RFs: ', t1, 'sec')\n",
        "print('wall time of building RFs: ', t1/60, 'min')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_dl7a1-2tQnD"
      },
      "source": [
        "def RF_Best_Model(analysis_objects, verbose=False):\n",
        "    \n",
        "    fig, ax = plt.subplots(2, 1, figsize=(15,22 ))\n",
        "    \n",
        "    # Setup comparison plot\n",
        "    ax[0].plot([0, 1], [0, 1], color=\"green\", linestyle='-.', label=\"Random Guess\")\n",
        "    ax[0].set_xlabel(\"False Positive\")\n",
        "    ax[0].set_ylabel(\"True Positive\")\n",
        "    ax[0].set_title(\"ROC Curves\")\n",
        "    \n",
        "    best_model = None\n",
        "    max_auc = 0\n",
        "    auc_values = []\n",
        "        \n",
        "    for key in analysis_objects:\n",
        "        analysis_objects[key].Compute_ROC_Curve()\n",
        "        ax[0].plot(analysis_objects[key].fpr, analysis_objects[key].tpr)\n",
        "        roc_auc = analysis_objects[key].AUC_score\n",
        "        auc_values.append(roc_auc)\n",
        "        if roc_auc > max_auc:\n",
        "            best_model = analysis_objects[key]\n",
        "            max_auc = roc_auc\n",
        "        \n",
        "        if verbose:\n",
        "            print(key, \"ROC AUC = \", roc_auc)\n",
        "            \n",
        "    \n",
        "    # Setup best model plot\n",
        "    best_model.Plot_ROC_Curve(axes=ax[1])\n",
        " \n",
        "    hist_ax = fig.add_axes([0.22, 0.23, 0.1, 0.13])\n",
        "    hist_ax.hist(auc_values, bins=len(analysis_objects))\n",
        "    hist_ax.set_yticks([])\n",
        " \n",
        "    return best_model"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qmEsSP7DIGRq"
      },
      "source": [
        "rf.predict_proba(X_test) \n",
        "print(best_RF)\n",
        "print(pred)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yFgg9vQoIGaC"
      },
      "source": [
        "print('random forest accuracy is ',rf.predict(X_test))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2R2QFreSk2Pm"
      },
      "source": [
        "from  sklearn.metrics import mean_squared_error"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4vKlgiMD4Smz"
      },
      "source": [
        "# mse = mean_squared_error(actual, predicted)\n",
        "predicted = rf.predict(X_test)\n",
        "mse = mean_squared_error(Y_test, predicted)\n",
        "rmse = math.sqrt(mse)\n",
        "print('root mean square error for Random Forest',100*max(0,rmse)) \n",
        "print('accuracy of random forest',rf.score(X_test,Y_test))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZyYQFhQmIGrk"
      },
      "source": [
        "def evaluate_model(model, X, y):\n",
        "    # define the evaluation procedure\n",
        "    cv = RepeatedStratifiedKFold(n_splits=10, n_repeats=3, random_state=1)\n",
        "    # evaluate the model and collect the results\n",
        "    scores = cross_val_score(model, X, y, scoring='accuracy', cv=cv, n_jobs=-1)\n",
        "    return scores"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6AWSaamB4eoq"
      },
      "source": [
        "test_scores = []\n",
        "test_yhat = rf.predict(X_test)\n",
        "test_acc = accuracy_score(Y_test, test_yhat)\n",
        "test_scores.append(test_acc)\n",
        "# summarize progress\n",
        "print('test: %.3f' % (test_acc))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TAj9AcIA4j_W"
      },
      "source": [
        "best_randomforest = RF_Best_Model(best_RF, verbose=True)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "B7frXFASG6y-"
      },
      "source": [
        "best_randomforest.model"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RkVVTZNj3pMZ"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "iSti0yravsJA"
      },
      "source": [
        "## Results\n",
        "True positive rate and false positive rate and also threshhold for the best RF have been driven. The results are significant that the model is eligible for the problem; discrimination of neutrinos versus background. "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "82FHVDK34rLY"
      },
      "source": [
        "best_randomforest.Display_Results()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_SMa-1OZx0jO"
      },
      "source": [
        "### Score of the best model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tz4RQzgrGojK"
      },
      "source": [
        "Y_pred_test = rf.predict(X_test)\n",
        "accuracy_score(Y_test, Y_pred_test)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "N-_yKuMq4v9t"
      },
      "source": [
        "target_names = ['cos', 'nu']\n",
        "print(classification_report(Y_test, Y_pred_test, target_names=target_names))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Uq2vXeEXxu4H"
      },
      "source": [
        "## Confusion Matrix"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XY6PnwZlGoqE"
      },
      "source": [
        "con_mat = confusion_matrix(Y_test, Y_pred_test)\n",
        "print(con_mat)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EWHD6NMu4z4I"
      },
      "source": [
        "con_mat = con_mat.astype('float') / con_mat.sum(axis=1)[:, np.newaxis]\n",
        "plt.figure(figsize=(8,6))\n",
        "sns.set(font_scale=1.4)\n",
        "sns.heatmap(con_mat, annot=True, annot_kws={'size':16}, cmap='plasma', linewidths=0.2)\n",
        "class_names = ['cos', 'nu']\n",
        "tick_marks = np.arange(len(class_names))\n",
        "tick_marks2 = tick_marks + 0.5\n",
        "plt.xticks(tick_marks, class_names)\n",
        "plt.yticks(tick_marks2, class_names)\n",
        "plt.xlabel('Predicted label')\n",
        "plt.ylabel('True label')\n",
        "plt.title('Confusion Matrix for RF')\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "K2DrMBDZJOuW"
      },
      "source": [
        "from math import *"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1c_x_WVBCRlp"
      },
      "source": [
        "TN = 932407; FP = 1662; FN = 8150; TP = 151769\n",
        "\n",
        "\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6c8Om6HPCRlq"
      },
      "source": [
        "TN = 932407; FP = 1662; FN = 8150; TP = 151769\n",
        "print('true negative(TN):=>', TN , '||','false positive(FP) :=>', FP)\n",
        "print('================================================================')\n",
        "print('false negative(FN):=>', FN ,' ||','true positive(TP) :=>', TP)\n",
        "print('----------------------------------------------------------------')\n",
        "P = FN + TP; N = TN + FP\n",
        "print('P =>', P, 'N =>', N)\n",
        "print('----------------------------------------------------------------')\n",
        "TPR = TP/P; FNR = 1 - TPR\n",
        "TNR = TN/N; FPR = 1 - TNR\n",
        "PPV = TP/(TP+FP); FDR = 1 - PPV\n",
        "NPV = TN/(TN+FN); FOR = 1 - NPV\n",
        "TS = TP/(TP+FN+FP)\n",
        "ACC = (TP+TN)/(P+N)\n",
        "BA = (TPR+TNR)/2\n",
        "print('\\033[1m\\033[91mSensitivity\\033[1m\\033[91m or \\033[1m\\033[91mtrue positive rate(TPR)\\033[0m ==>', TPR)\n",
        "\n",
        "print('\\033[1m\\033[91mMISS RATE or false negative rate(FNR)\\033[0m ==>', FNR)\n",
        "print('_______________________________________________________________________________')\n",
        "print('\\033[1m\\033[91mspecificity, selectivity or true negative rate(TNR)\\033[0m ==>', TNR)\n",
        "\n",
        "print('\\033[1m\\033[91mfall-out or false positive rate (FPR)\\033[0m ==>', FPR)\n",
        "print('_______________________________________________________________________________')\n",
        "print('\\033[1m\\033[91mprecision or positive predictive value(PPV)\\033[0m ==>', PPV)\n",
        "\n",
        "print('\\033[1m\\033[91mfalse discovery rate (FDR)\\033[0m ==>', FDR)\n",
        "print('_______________________________________________________________________________')\n",
        "print('\\033[1m\\033[91mnegative predictive value(NPV)\\033[0m ==>', NPV)\n",
        "\n",
        "print('\\033[1m\\033[91mfalse omission rate (FOR)\\033[0m ==>', FOR)\n",
        "print('_______________________________________________________________________________')\n",
        "print('\\033[1m\\033[91mthreat score(TS) or critical success index(CSI)\\033[0m ==>', TS)\n",
        "\n",
        "print('/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\\\/\\//\\/\\/\\/\\/\\/\\/\\//\\/\\/\\//\\/\\/\\/\\/\\/\\\\//')\n",
        "\n",
        "div = (sqrt(TPR*(-TNR+1)))+TNR-1\n",
        "divder = TPR+TNR-1\n",
        "PT = div/divder\n",
        "print('\\033[1m\\033[91mprevalence threshold(PT)\\033[0m ==>', PT)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_Vf_aiI0CRlr"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "M7GCqQnCCRlt"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FsC4ZyFox-qw"
      },
      "source": [
        "# The outcomes"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AV1wISWXzgMP"
      },
      "source": [
        "## Survey the model\n",
        "In this stage, we studied the test data, findinghow the model could predict the data. The tast data has been already provided. Here Y_pred_test represented the model prediction and Y_test means true label of test data. The subtraction of these two labels shows if the model classified the input data correctly or not."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3cHW_4hDGo2X"
      },
      "source": [
        "er = Y_test - Y_pred_test\n",
        "print(er)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9tV0IlRZ0VVP"
      },
      "source": [
        "### Appropriate Prediction Subset of Data "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RRpfpB2SjyOU"
      },
      "source": [
        "corrcases = er[er==0]\n",
        "corrcases = corrcases.sort_index()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-z2aK6aAjypr"
      },
      "source": [
        "corrcases_fr = corrcases.to_frame()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SeQ55kQmjywz"
      },
      "source": [
        "corr_inx = corrcases_fr.index"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PVd0HiLHjy54"
      },
      "source": [
        "corr_df = data.loc[corr_inx]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5qEnZ0huj69J"
      },
      "source": [
        "corr_nue = corr_df.loc[corr_df['label']==1]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "c_Na12r1j7Hc"
      },
      "source": [
        "no_corrnue = len(corr_nue)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "P2_Zv_3ej7QH"
      },
      "source": [
        "corr_cosmic = corr_df.loc[corr_df['label']==0]\n",
        "corr_cosmic"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KdgJ60RQj7ka"
      },
      "source": [
        "no_corr_cosmic = len(corr_cosmic)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "byrT7q6RjzCD"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zJiKNzW5jzKb"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XXkUSXvp1Mcn"
      },
      "source": [
        "### nue files which have been predicted as cosmics"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hl761f7SFYf2"
      },
      "source": [
        "ntcases = er[er==1]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BDpFlpv0FYm2"
      },
      "source": [
        "print('false negative =')\n",
        "ntcases  = ntcases.sort_index()\n",
        "print(ntcases)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "C7jy_iDWFYvo"
      },
      "source": [
        "ntcases_df = ntcases.to_frame()\n",
        "ntcases_df"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WjtE3CDle-52"
      },
      "source": [
        "inx_falsenegative = ntcases_df.index"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ehCu00nmFY7w"
      },
      "source": [
        "falsenegative_df = data.loc[inx_falsenegative]\n",
        "print('missclassified signals predicted as bacground:=:', falsenegative_df)\n",
        "falsenegative_df"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SeZmURjoXz8s"
      },
      "source": [
        "print('Whole missclassified signals predicted as bacground:=:')\n",
        "print_full(falsenegative_df)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sJ6NHDPzEs_M"
      },
      "source": [
        "falsenegativedata = falsenegative_df.drop(['label', 'hit_tpc',\t'hit_plane'], axis=1)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4TpziOKmE56K"
      },
      "source": [
        "https://colab.research.google.com/drive/1N4waqDsOYKu_Uf69-78rpK8CzYLQQuzt?usp=sharing"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "aVnVMgfbjhU9"
      },
      "source": [
        "##wrong presiction over nue:\n",
        "\n",
        "The plot below reports the percentage of wrong prediction of \"part\" of an event(entry). Our input nue data has 100*19 = 1900 entrys. As mentioned before, some of subentrys of each entry have been selected for tranning data. Hence, test data set aka unseen data containes some of full event."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "oajj0rHiUbmc"
      },
      "source": [
        "#fig = px.scatter(wrationue, x = 'entry', y = 'percentage of misclassified subentry', title= 'misclassified nue')\n",
        "#fig.show()\n",
        "https://colab.research.google.com/drive/1N4waqDsOYKu_Uf69-78rpK8CzYLQQuzt#scrollTo=oajj0rHiUbmc&line=1&uniqifier=1"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kbocEWsUjgTE"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "u4efC78-1hUF"
      },
      "source": [
        "### cosmic files that were predicted as nues"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dG23OagnFZIa"
      },
      "source": [
        "pfcases = er[er==-1]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9ZYMPPswFZPg"
      },
      "source": [
        "print('false positive =')\n",
        "pfcases  = pfcases.sort_index()\n",
        "print(pfcases)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9uGA701hFZWo"
      },
      "source": [
        "positivetrue_df = pfcases.to_frame()\n",
        "inx_positiveture = positivetrue_df.index"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PhUUgZn8HdVM"
      },
      "source": [
        "positivetrue_df.index"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PhV40QeoHdcf"
      },
      "source": [
        "ptdataframe = data.loc[positivetrue_df.index]\n",
        "print('missclassified bacground predicted as signals:=:', ptdataframe)\n",
        "ptdataframe"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EA15LiYLX3O6"
      },
      "source": [
        "print('Whole missclassified bacgrounds predicted as signals:=:')\n",
        "print_full(ptdataframe)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NHeJgIEtCsQ6"
      },
      "source": [
        "df_ptdataframe = ptdataframe.drop(['label', 'hit_tpc',\t'hit_plane'], axis=1)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VRh8hi86E1NV"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "efkEws9rHdjs"
      },
      "source": [
        "no_data = len(data.index.get_level_values('entry').unique().to_numpy()) + len(df_in0_2.index.get_level_values('entry').unique().to_numpy())\n",
        "no_falsenegative = len(ntcases_df.index.get_level_values('entry').unique().to_numpy())\n",
        "no_falspositve = len(positivetrue_df.index.get_level_values('entry').unique().to_numpy())\n",
        "no_true = no_data - no_falsenegative"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SukQG7mfQMfs"
      },
      "source": [
        "falsenegative_df.info()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "f3UlsJP-1yHL"
      },
      "source": [
        "### The distribution of missclassified subentrys through entrys for cosmic files"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "p2VRyyHKsTap"
      },
      "source": [
        "Fposdis = []\n",
        "for i in range(len(ptdataframe)):\n",
        "  xcv = ptdataframe.iloc[ptdataframe.index.get_level_values('entry') == i]\n",
        "  lxcv = len(xcv)\n",
        "  print(lxcv)\n",
        "  Fposdis.append(lxcv)\n",
        "\n",
        "#FNegdis_df.plot(figsize=(18,16), title='False Positive cases')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4p494DrJsTyQ"
      },
      "source": [
        "Fposdis_df = pd.DataFrame(Fposdis)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "B0ZaWldDsYd_"
      },
      "source": [
        "print_full(Fposdis_df)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "STg-c3zzQDnJ"
      },
      "source": [
        "Fposdis_df_vi = Fposdis_df.T"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4dlfCmkH2mWi"
      },
      "source": [
        "###The distribution of missclassified subentrys through entrys for cosmic files"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sKPO2tTCstvR"
      },
      "source": [
        "FHNegdis_df = pd.DataFrame(FHNegdis)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ek_C0Z1lsxsK"
      },
      "source": [
        "print_full(FHNegdis_df)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yOWql92vR2RW"
      },
      "source": [
        "https://colab.research.google.com/drive/1N4waqDsOYKu_Uf69-78rpK8CzYLQQuzt?usp=sharing"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5GkTJPAmle0E"
      },
      "source": [
        "##WRONG PREDICTION COSMIC EVENTS"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WSJmHRNCUbmN"
      },
      "source": [
        "fig = px.scatter(wrationue, x = 'entry', y = 'percentage of misclassified subentry', title= 'misclassified cosmic')\n",
        "fig.show()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "F0ERdZObomNx"
      },
      "source": [
        "###Score of the model\n",
        "\n",
        "Based on the results came out from the two graphs which indicated the percentage of wrong subentrys for each enty, one could extracte the score and other statistical detailes of the best RF"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-WkYi1lsoDBl"
      },
      "source": [
        "#model.info in entry level\n",
        "TN = 9261; FP = 231; FN = 738; TP = 1162\n",
        "print('true negative(TN):=>', TN , '||','false positive(FP) :=>', FP)\n",
        "print('================================================================')\n",
        "print('false negative(FN):=>', FN ,' ||','true positive(TP) :=>', TP)\n",
        "print('----------------------------------------------------------------')\n",
        "P = FN + TP; N = TN + FP\n",
        "print('P =>', P, 'N =>', N)\n",
        "print('----------------------------------------------------------------')\n",
        "TPR = TP/P; FNR = 1 - TPR\n",
        "TNR = TN/N; FPR = 1 - TNR\n",
        "PPV = TP/(TP+FP); FDR = 1 - PPV\n",
        "NPV = TN/(TN+FN); FOR = 1 - NPV\n",
        "TS = TP/(TP+FN+FP)\n",
        "ACC = (TP+TN)/(P+N)\n",
        "BA = (TPR+TNR)/2\n",
        "print('\\033[1m\\033[91mSensitivity\\033[1m\\033[91m or \\033[1m\\033[91mtrue positive rate(TPR)\\033[0m ==>', TPR)\n",
        "\n",
        "print('\\033[1m\\033[91mMISS RATE or false negative rate(FNR)\\033[0m ==>', FNR)\n",
        "print('_______________________________________________________________________________')\n",
        "print('\\033[1m\\033[91mspecificity, selectivity or true negative rate(TNR)\\033[0m ==>', TNR)\n",
        "\n",
        "print('\\033[1m\\033[91mfall-out or false positive rate (FPR)\\033[0m ==>', FPR)\n",
        "print('_______________________________________________________________________________')\n",
        "print('\\033[1m\\033[91mprecision or positive predictive value(PPV)\\033[0m ==>', PPV)\n",
        "\n",
        "print('\\033[1m\\033[91mfalse discovery rate (FDR)\\033[0m ==>', FDR)\n",
        "print('_______________________________________________________________________________')\n",
        "print('\\033[1m\\033[91mnegative predictive value(NPV)\\033[0m ==>', NPV)\n",
        "\n",
        "print('\\033[1m\\033[91mfalse omission rate (FOR)\\033[0m ==>', FOR)\n",
        "print('_______________________________________________________________________________')\n",
        "print('\\033[1m\\033[91mthreat score(TS) or critical success index(CSI)\\033[0m ==>', TS)\n",
        "\n",
        "print('/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\\\/\\//\\/\\/\\/\\/\\/\\/\\//\\/\\/\\//\\/\\/\\/\\/\\/\\\\//')\n",
        "\n",
        "div = (sqrt(TPR*(-TNR+1)))+TNR-1\n",
        "divder = TPR+TNR-1\n",
        "PT = div/divder\n",
        "print('\\033[1m\\033[91mprevalence threshold(PT)\\033[0m ==>', PT)\n",
        "ACC = (TP+TN)/(P+N)\n",
        "print('===================================================================================')\n",
        "print('===================================================================================')\n",
        "print('\\033[1m\\033[91mAccuracy of the model - entry level(ACC)\\033[0m ==>', ACC)\n",
        "print('===================================================================================')\n",
        "print('===================================================================================')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "brP3edCx26nZ"
      },
      "source": [
        "## plot of correct and missclassified subentrys for an entry of nue files in the same canvas"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fXT0cK5zb4CZ"
      },
      "source": [
        "for i, j in zip(range(no_corrnue), range(no_falsenegative)):\n",
        "\n",
        "  cordata = corr_nue.iloc[corr_nue.index.get_level_values('entry') == i]\n",
        "  cp1 = cordata['hit_plane'] ==1\n",
        "  cii = cordata['hit_plane'] ==2\n",
        "  czo = cordata['hit_tpc'] ==0\n",
        "  cun = cordata['hit_tpc'] ==1\n",
        "  plt.title('MISSCLASSIFIED SIGNALS VERSUS CORRECT ONES', fontsize=18)\n",
        "  plt.figure(figsize=(15,18))\n",
        "  plt.xlabel('hit_wire', fontsize=18)\n",
        "  plt.ylabel('hit_peaktime', fontsize=18)\n",
        "  plt.scatter(cordata['hit_wire'],cordata['hit_peakT'], color = 'black', marker = '.')\n",
        "  plt.scatter(cordata['hit_wire'].loc[cii&czo],cordata['hit_peakT'].loc[cii&czo], color = 'blue', marker = '.')\n",
        "  plt.scatter(cordata['hit_wire'].loc[cp1&cun],cordata['hit_peakT'].loc[cp1&cun], color = 'green', marker = '+')\n",
        "  indata = falsenegative_df.iloc[falsenegative_df.index.get_level_values('entry') == j]\n",
        "  p1 = indata['hit_plane'] ==1 \n",
        "  ii = indata['hit_plane'] ==2\n",
        "  zo = indata['hit_tpc'] ==0\n",
        "  un = indata['hit_tpc'] ==1\n",
        "  plt.scatter(indata['hit_wire'].loc[ii&zo],indata['hit_peakT'].loc[ii&zo], color = 'black', marker = '*')\n",
        "  plt.scatter(indata['hit_wire'].loc[p1&un],indata['hit_peakT'].loc[p1&un], color = 'red', marker = '+')\n",
        "  ll = [ 'corr_plane=2, corr_tpc=0','corr_plane=1, corr_tpc=1', 'miss_plane=2, miss_tpc=0','miss_plane=1, miss_tpc=1']\n",
        "  plt.legend(ll, loc=\"upper left\", title=\"Classes\") \n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Q5J1Z0ThiSHH"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QyvlvAOm3Q7K"
      },
      "source": [
        "## plot of correct and missclassified subentrys for an entry of cosmic filesin the same canvas"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GpdzL8elhQYw"
      },
      "source": [
        "for i, j in zip(range(no_corr_cosmic), range(no_falspositve)):\n",
        "\n",
        "  cordata = corr_cosmic.iloc[corr_cosmic.index.get_level_values('entry') == i]\n",
        "  cp1 = cordata['hit_plane'] ==1\n",
        "  cii = cordata['hit_plane'] ==2\n",
        "  czo = cordata['hit_tpc'] ==0\n",
        "  cun = cordata['hit_tpc'] ==1\n",
        "  plt.title('MISSCLASSIFIED COSMICS VERSUS CORRECT ONES', fontsize=18)\n",
        "  plt.figure(figsize=(15,18))\n",
        "  plt.xlabel('hit_wire', fontsize=18)\n",
        "  plt.ylabel('hit_peaktime', fontsize=18)\n",
        "  plt.scatter(cordata['hit_wire'].loc[cii&czo],cordata['hit_peakT'].loc[cii&czo], color = 'green', marker = '.')\n",
        "  plt.scatter(cordata['hit_wire'].loc[cp1&cun],cordata['hit_peakT'].loc[cp1&cun], color = 'blue', marker = '.')\n",
        "\n",
        "  indata = ptdataframe.iloc[ptdataframe.index.get_level_values('entry') == j]\n",
        "  p1 = indata['hit_plane'] ==1 \n",
        "  ii = indata['hit_plane'] ==2\n",
        "  zo = indata['hit_tpc'] ==0\n",
        "  un = indata['hit_tpc'] ==1\n",
        "  plt.scatter(indata['hit_wire'].loc[ii&zo],indata['hit_peakT'].loc[ii&zo], color = 'black', marker = '*')\n",
        "  plt.scatter(indata['hit_wire'].loc[p1&un],indata['hit_peakT'].loc[p1&un], color = 'red', marker = '*')\n",
        "  ll = [ 'corr_plane=2, corr_tpc=0','cos_plane=1, cos_tpc=1', 'miss_plane=2, miss_tpc=0','miss_plane=1, miss_tpc=1']\n",
        "  plt.legend(ll, loc=\"upper left\", title=\"Classes\") \n",
        "\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LgZPVWgWHdrp"
      },
      "source": [
        "print('\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/\\/')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-LJ6IJ6A0EzH"
      },
      "source": [
        "#OUTCOMES\n",
        "\n",
        "Based on results, the best RF model recorded the better efficiency to classify nue and cosmic for subentry level. The reason behind that refers to this subject that the samples or training data are single hit. Despite of that, the model turned over data from Collection (hit_plane = 2, hit_tpc = 0) granted the best performance, while the algorithm for induction 2 can predict the lower amount of single hits and also event. This might refer to this fact that Collection of ICARUS T600 has the highest resolution.\n",
        "In all of cases, TNR is significantly more than TPR which tells us that the algorithm can recognize cosmic rays. It is also predictable as the size of cosmic 10 times higher than nue.\n",
        "Consequently the performance of the model (specially TPR) increased when the input data came from collection with tpc = 0 in both entry and subentry level. Although the percentage of wrong predicted entry "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lRjivnJs0b51"
      },
      "source": [
        "#Further down the road\n",
        "\n",
        "\n",
        "\n",
        "*   Data with overlapped cosmic and neutrino hits\n",
        "*   Application to real data with lower signal-to-noise ratio\n",
        "*   Using real data with the current RF model and tuning the parameters of the model to increase the accurac\n",
        "\n",
        "As mentioned before, the input data was true or simulated data. Therefore they are pretty clean without formidable noise. The next study could be including the reconstructed data with lower signal to noise ratio. For stepping on t his real life data, it is required to revision and reassessment of input data. The other features could be added to them like the deposited energy of the hits. Also the model might be eligible (with some manipulations) for  screening different types of neutrons.\n",
        "\n",
        "\n"
      ]
    }
  ]
}