STOCK_PREDICTION_LSTM_RNN.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "STOCK_PREDICTION_LSTM_RNN.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "QpIxV1wusR8G"
      },
      "source": [
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n"
      ],
      "execution_count": 20,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 203
        },
        "id": "9Y7oazLaECEz",
        "outputId": "7b59d01d-5be2-49b6-c1da-a6df185eafea"
      },
      "source": [
        "dataset= pd.read_csv('Google_Stock_Price_Train.csv')\n",
        "training_set = dataset.iloc[:, 1:2]\n",
        "training_set.head()"
      ],
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Open</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>325.25</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>331.27</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>329.83</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>328.34</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>322.04</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "     Open\n",
              "0  325.25\n",
              "1  331.27\n",
              "2  329.83\n",
              "3  328.34\n",
              "4  322.04"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 21
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FYiYP571EYMy"
      },
      "source": [
        "#feature scaling\n",
        "from sklearn.preprocessing import MinMaxScaler#\n",
        "\n",
        "sc = MinMaxScaler(feature_range=(0,1))\n",
        "training_set_scaled = sc.fit_transform(training_set)"
      ],
      "execution_count": 22,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4WnAmP9XE0R-"
      },
      "source": [
        "X_train=[]\n",
        "Y_train=[]\n",
        "for i in range(60,1258):\n",
        "  X_train.append(training_set_scaled[i-60:i,0])\n",
        "  Y_train.append(training_set_scaled[i,0])\n",
        "\n",
        "X_train,Y_train = np.array(X_train), np.array(Y_train)\n",
        "#reshape array\n",
        "X_train= np.reshape(X_train,(X_train.shape[0], X_train.shape[1],1))"
      ],
      "execution_count": 23,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6r6lZT6vGNeI"
      },
      "source": [
        "from keras.models import Sequential\n",
        "from  keras.layers import Dense ,LSTM , Dropout"
      ],
      "execution_count": 24,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Kfm5yhemGr3Q"
      },
      "source": [
        "regressor = Sequential()"
      ],
      "execution_count": 25,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QeGTd67ZNaWK"
      },
      "source": [
        "#add lstm layers and dropout regularization\n",
        "regressor.add(LSTM(units=50, return_sequences=True , input_shape=(X_train.shape[1],1)))\n",
        "regressor.add(Dropout(0.2))"
      ],
      "execution_count": 26,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GK7qAxIJR0bD"
      },
      "source": [
        "regressor.add(LSTM(units=50, return_sequences=True  ))\n",
        "regressor.add(Dropout(0.2))"
      ],
      "execution_count": 27,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ss8ScS2lbARM"
      },
      "source": [
        "regressor.add(LSTM(units=50, return_sequences=True  ))\n",
        "regressor.add(Dropout(0.2))"
      ],
      "execution_count": 28,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "n2-zu2CAbCq2"
      },
      "source": [
        "regressor.add(LSTM(units=50  ))\n",
        "regressor.add(Dropout(0.2))"
      ],
      "execution_count": 29,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nDpC6Vy4bIZy"
      },
      "source": [
        "regressor.add(Dense(units=1))#output layer"
      ],
      "execution_count": 30,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "k6d93T20bOt3"
      },
      "source": [
        "#compile rnn\n",
        "regressor.compile(optimizer='adam', loss='mean_squared_error')"
      ],
      "execution_count": 31,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "a4JQAv0Qbzjk"
      },
      "source": [
        "regressor.fit(X_train,Y_train,epochs=100,batch_size=32)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GEdNjfvwcAkM"
      },
      "source": [
        "dataset_test=pd.read_csv('Google_Stock_Price_Test.csv')\n",
        "real_stock_price = dataset_test.iloc[:,1:2].values"
      ],
      "execution_count": 33,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ICuarM-pcWuI"
      },
      "source": [
        "dataset_total= pd.concat((dataset['Open'], dataset_test['Open']), axis=0)\n",
        "inputs = dataset_total[len(dataset_total)- len(dataset_test)-60:].values\n",
        "inputs =inputs.reshape(-1,1)\n",
        "inputs = sc.transform(inputs)\n",
        "X_test=[]\n",
        "for i in range(60,80):\n",
        "  X_test.append(inputs[i-60:i , 0])\n",
        "\n",
        "X_test = np.array(X_test)\n",
        "X_test =  np.reshape(X_test,(X_test.shape[0],X_test.shape[1],1))\n",
        "predicted_stock_price = regressor.predict(X_test)\n",
        "predicted_stock_price = sc.inverse_transform(predicted_stock_price)\n",
        " "
      ],
      "execution_count": 37,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "Jm7DvEYOG28j",
        "outputId": "146ce3f2-6a73-4223-ef51-f37a0fb4c36e"
      },
      "source": [
        "#lets visualize the results of predicted and real stock price\n",
        "plt.plot(real_stock_price, color='r', label='Real Google Stock Price')\n",
        "plt.plot(predicted_stock_price, color='b', label='Predicted Google Stock Price')\n",
        "\n",
        "plt.title('Google Stock Price Prediction')\n",
        "plt.xlabel('Time')\n",
        "plt.ylabel('Google Stock Price')\n",
        "plt.legend()\n",
        "plt.show()\n"
      ],
      "execution_count": 38,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}