Stoc_pred.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Stoc_pred.ipynb",
      "provenance": [],
      "authorship_tag": "ABX9TyPH9gifaaLJjZKVTImhx+cY",
      "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/feruxhi/Stock_Prediction/blob/main/Stoc_pred.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "iG6Q0lnPZnKw",
        "outputId": "16676f96-4d75-4b33-ee3f-530ffdab58f1"
      },
      "source": [
        "import numpy as np\r\n",
        "import pandas as pd\r\n",
        "import matplotlib.pyplot as plt\r\n",
        "from sklearn.metrics import mean_absolute_error as mae\r\n",
        "from sklearn.preprocessing import StandardScaler\r\n",
        "\r\n",
        "from keras.models import Model\r\n",
        "from keras.layers import Input, LSTM, GRU, SimpleRNN, Dense\r\n",
        "import keras.backend as K\r\n",
        "from keras.optimizers import SGD, Adam\r\n",
        "\r\n",
        "# get the data\r\n",
        "df = pd.read_csv('https://raw.githubusercontent.com/lazyprogrammer/machine_learning_examples/master/tf2.0/sbux.csv')\r\n",
        "\r\n",
        "df.head()\r\n",
        "# get the close price\r\n",
        "series = df['close'].values.reshape(-1, 1)\r\n",
        "# Normalize the data\r\n",
        "# Note: I didn't think about where the true boundary is, this is just approx.\r\n",
        "scaler = StandardScaler()\r\n",
        "scaler.fit(series[:len(series) // 2])\r\n",
        "series = scaler.transform(series).flatten()\r\n",
        "# plot it\r\n",
        "plt.plot(series)\r\n",
        "plt.show()\r\n",
        "\r\n",
        "### build the dataset\r\n",
        "# let's see if we can use T past values to predict the next value\r\n",
        "T = 10\r\n",
        "D = 1\r\n",
        "X = []\r\n",
        "Y = []\r\n",
        "for t in range(len(series) - T):\r\n",
        "  x = series[t:t+T]\r\n",
        "  X.append(x)\r\n",
        "  y = series[t+T]\r\n",
        "  Y.append(y)\r\n",
        "\r\n",
        "X = np.array(X).reshape(-1, T, 1)\r\n",
        "Y = np.array(Y)\r\n",
        "N = len(X)\r\n",
        "print(\"X.shape\", X.shape, \"Y.shape\", Y.shape)\r\n",
        "\r\n",
        "# split into train and test\r\n",
        "Xtrain, Ytrain = X[:-N//2], Y[:-N//2]\r\n",
        "Xtest, Ytest = X[-N//2:], Y[-N//2:]\r\n",
        "\r\n",
        "### try autoregressive RNN model\r\n",
        "i = Input(shape=(T, 1))\r\n",
        "\r\n",
        "# multiple LSTMs\r\n",
        "# x = LSTM(5, return_sequences=True)(i)\r\n",
        "# x = LSTM(5, return_sequences=True)(x)\r\n",
        "# x = LSTM(5)(x)\r\n",
        "\r\n",
        "# one LSTM\r\n",
        "x = LSTM(5)(i)\r\n",
        "\r\n",
        "x = Dense(1)(x)\r\n",
        "model = Model(i, x)\r\n",
        "model.compile(\r\n",
        "  loss='mse',\r\n",
        "  optimizer=Adam(lr=0.01),\r\n",
        ")\r\n",
        "# train the RNN\r\n",
        "r = model.fit(\r\n",
        "  Xtrain, Ytrain,\r\n",
        "  epochs=80,\r\n",
        "  validation_data=(Xtest, Ytest),\r\n",
        ")\r\n",
        "# Plot loss per iteration\r\n",
        "plt.plot(r.history['loss'], label='loss')\r\n",
        "plt.plot(r.history['val_loss'], label='val_loss')\r\n",
        "plt.legend()\r\n",
        "plt.show()\r\n",
        "\r\n",
        "# Make predictions\r\n",
        "Ptrain = model.predict(Xtrain)\r\n",
        "Ptest = model.predict(Xtest)\r\n",
        "# inverse transform\r\n",
        "Ytrain2 = scaler.inverse_transform(Ytrain.reshape(-1, 1)).flatten()\r\n",
        "Ytest2 = scaler.inverse_transform(Ytest.reshape(-1, 1)).flatten()\r\n",
        "Ptrain2 = scaler.inverse_transform(Ptrain.reshape(-1, 1)).flatten()\r\n",
        "Ptest2 = scaler.inverse_transform(Ptest.reshape(-1, 1)).flatten()\r\n",
        "# Print the MAE\r\n",
        "print(\"lstm train mae:\", mae(Ytrain2, Ptrain2))\r\n",
        "print(\"lstm test mae:\", mae(Ytest2, Ptest2))\r\n",
        "# wrong forecast\r\n",
        "plt.plot(Ytest, label='target')\r\n",
        "plt.plot(Ptest, label='prediction')\r\n",
        "plt.legend()\r\n",
        "plt.title(\"Wrong forecast\")\r\n",
        "plt.show()\r\n",
        "# right forecast\r\n",
        "forecast = []\r\n",
        "input_ = Xtest[0]\r\n",
        "while len(forecast) < len(Ytest):\r\n",
        "  f = model.predict(input_.reshape(1, T, 1))[0,0]\r\n",
        "  forecast.append(f)\r\n",
        "\r\n",
        "  # make a new input with the latest forecast\r\n",
        "  input_ = np.roll(input_, -1)\r\n",
        "  input_[-1] = f\r\n",
        "plt.plot(Ytest, label='target')\r\n",
        "plt.plot(forecast, label='prediction')\r\n",
        "plt.legend()\r\n",
        "plt.title(\"Right forecast\")\r\n",
        "plt.show()"
      ],
      "execution_count": null,
      "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"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "X.shape (1249, 10, 1) Y.shape (1249,)\n",
            "Epoch 1/80\n",
            "20/20 [==============================] - 3s 44ms/step - loss: 1.0229 - val_loss: 3.3136\n",
            "Epoch 2/80\n",
            "20/20 [==============================] - 0s 8ms/step - loss: 0.1974 - val_loss: 0.8268\n",
            "Epoch 3/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0778 - val_loss: 0.5326\n",
            "Epoch 4/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0377 - val_loss: 0.3216\n",
            "Epoch 5/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0182 - val_loss: 0.1838\n",
            "Epoch 6/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0113 - val_loss: 0.1047\n",
            "Epoch 7/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0085 - val_loss: 0.0893\n",
            "Epoch 8/80\n",
            "20/20 [==============================] - 0s 8ms/step - loss: 0.0076 - val_loss: 0.0742\n",
            "Epoch 9/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0076 - val_loss: 0.0676\n",
            "Epoch 10/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0068 - val_loss: 0.0566\n",
            "Epoch 11/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0074 - val_loss: 0.0542\n",
            "Epoch 12/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0061 - val_loss: 0.0520\n",
            "Epoch 13/80\n",
            "20/20 [==============================] - 0s 8ms/step - loss: 0.0059 - val_loss: 0.0499\n",
            "Epoch 14/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0060 - val_loss: 0.0489\n",
            "Epoch 15/80\n",
            "20/20 [==============================] - 0s 8ms/step - loss: 0.0064 - val_loss: 0.0469\n",
            "Epoch 16/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0062 - val_loss: 0.0464\n",
            "Epoch 17/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0062 - val_loss: 0.0472\n",
            "Epoch 18/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0065 - val_loss: 0.0450\n",
            "Epoch 19/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0067 - val_loss: 0.0432\n",
            "Epoch 20/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0066 - val_loss: 0.0429\n",
            "Epoch 21/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0063 - val_loss: 0.0411\n",
            "Epoch 22/80\n",
            "20/20 [==============================] - 0s 8ms/step - loss: 0.0056 - val_loss: 0.0403\n",
            "Epoch 23/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0070 - val_loss: 0.0392\n",
            "Epoch 24/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0053 - val_loss: 0.0393\n",
            "Epoch 25/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0053 - val_loss: 0.0389\n",
            "Epoch 26/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0059 - val_loss: 0.0354\n",
            "Epoch 27/80\n",
            "20/20 [==============================] - 0s 6ms/step - loss: 0.0060 - val_loss: 0.0361\n",
            "Epoch 28/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0061 - val_loss: 0.0343\n",
            "Epoch 29/80\n",
            "20/20 [==============================] - 0s 8ms/step - loss: 0.0057 - val_loss: 0.0384\n",
            "Epoch 30/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0064 - val_loss: 0.0365\n",
            "Epoch 31/80\n",
            "20/20 [==============================] - 0s 6ms/step - loss: 0.0059 - val_loss: 0.0326\n",
            "Epoch 32/80\n",
            "20/20 [==============================] - 0s 6ms/step - loss: 0.0069 - val_loss: 0.0327\n",
            "Epoch 33/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0063 - val_loss: 0.0314\n",
            "Epoch 34/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0061 - val_loss: 0.0351\n",
            "Epoch 35/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0055 - val_loss: 0.0306\n",
            "Epoch 36/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0048 - val_loss: 0.0299\n",
            "Epoch 37/80\n",
            "20/20 [==============================] - 0s 8ms/step - loss: 0.0060 - val_loss: 0.0291\n",
            "Epoch 38/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0053 - val_loss: 0.0286\n",
            "Epoch 39/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0047 - val_loss: 0.0303\n",
            "Epoch 40/80\n",
            "20/20 [==============================] - 0s 13ms/step - loss: 0.0058 - val_loss: 0.0308\n",
            "Epoch 41/80\n",
            "20/20 [==============================] - 0s 8ms/step - loss: 0.0051 - val_loss: 0.0276\n",
            "Epoch 42/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0052 - val_loss: 0.0294\n",
            "Epoch 43/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0055 - val_loss: 0.0274\n",
            "Epoch 44/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0046 - val_loss: 0.0276\n",
            "Epoch 45/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0054 - val_loss: 0.0260\n",
            "Epoch 46/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0050 - val_loss: 0.0256\n",
            "Epoch 47/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0053 - val_loss: 0.0262\n",
            "Epoch 48/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0049 - val_loss: 0.0253\n",
            "Epoch 49/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0049 - val_loss: 0.0244\n",
            "Epoch 50/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0057 - val_loss: 0.0243\n",
            "Epoch 51/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0052 - val_loss: 0.0243\n",
            "Epoch 52/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0051 - val_loss: 0.0242\n",
            "Epoch 53/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0044 - val_loss: 0.0238\n",
            "Epoch 54/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0052 - val_loss: 0.0240\n",
            "Epoch 55/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0051 - val_loss: 0.0286\n",
            "Epoch 56/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0048 - val_loss: 0.0225\n",
            "Epoch 57/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0054 - val_loss: 0.0226\n",
            "Epoch 58/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0051 - val_loss: 0.0222\n",
            "Epoch 59/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0048 - val_loss: 0.0229\n",
            "Epoch 60/80\n",
            "20/20 [==============================] - 0s 8ms/step - loss: 0.0059 - val_loss: 0.0218\n",
            "Epoch 61/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0055 - val_loss: 0.0227\n",
            "Epoch 62/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0055 - val_loss: 0.0217\n",
            "Epoch 63/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0050 - val_loss: 0.0211\n",
            "Epoch 64/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0056 - val_loss: 0.0227\n",
            "Epoch 65/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0057 - val_loss: 0.0217\n",
            "Epoch 66/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0049 - val_loss: 0.0215\n",
            "Epoch 67/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0055 - val_loss: 0.0211\n",
            "Epoch 68/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0050 - val_loss: 0.0236\n",
            "Epoch 69/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0049 - val_loss: 0.0275\n",
            "Epoch 70/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0060 - val_loss: 0.0285\n",
            "Epoch 71/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0051 - val_loss: 0.0203\n",
            "Epoch 72/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0051 - val_loss: 0.0200\n",
            "Epoch 73/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0055 - val_loss: 0.0201\n",
            "Epoch 74/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0056 - val_loss: 0.0210\n",
            "Epoch 75/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0047 - val_loss: 0.0209\n",
            "Epoch 76/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0053 - val_loss: 0.0200\n",
            "Epoch 77/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0046 - val_loss: 0.0190\n",
            "Epoch 78/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0064 - val_loss: 0.0213\n",
            "Epoch 79/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0048 - val_loss: 0.0218\n",
            "Epoch 80/80\n",
            "20/20 [==============================] - 0s 7ms/step - loss: 0.0051 - val_loss: 0.0275\n"
          ],
          "name": "stdout"
        },
        {
          "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"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "lstm train mae: 0.3792973873431868\n",
            "lstm test mae: 0.8871519582519529\n"
          ],
          "name": "stdout"
        },
        {
          "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"
          }
        },
        {
          "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"
          }
        }
      ]
    }
  ]
}