Developing_LSTM_Models_For_Univariate_Time_Series_Forecasting.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": " Developing LSTM Models For Univariate Time Series Forecasting.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "authorship_tag": "ABX9TyP1tVGd68GwTSpj4LKG4T9G",
      "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/arjun713/Automatic-leaf-infection-identifier/blob/master/Developing_LSTM_Models_For_Univariate_Time_Series_Forecasting.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FH4DBaTv_CVU"
      },
      "source": [
        "# univariate lstm example\n",
        "import numpy as np\n",
        "from tensorflow.keras.models import Sequential\n",
        "from tensorflow.keras.layers import LSTM\n",
        "from tensorflow.keras.layers import Dense\n",
        "from tensorflow.keras.layers import Flatten\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mh8f7iTvkleX"
      },
      "source": [
        "# preparing independent and dependent features\n",
        "def prepare_data(timeseries_data, n_features):\n",
        "\tX, y =[],[]\n",
        "\tfor i in range(len(timeseries_data)):\n",
        "\t\t# find the end of this pattern\n",
        "\t\tend_ix = i + n_features\n",
        "\t\t# check if we are beyond the sequence\n",
        "\t\tif end_ix > len(timeseries_data)-1:\n",
        "\t\t\tbreak\n",
        "\t\t# gather input and output parts of the pattern\n",
        "\t\tseq_x, seq_y = timeseries_data[i:end_ix], timeseries_data[end_ix]\n",
        "\t\tX.append(seq_x)\n",
        "\t\ty.append(seq_y)\n",
        "\treturn np.array(X), np.array(y)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "W-S5Gs7Qk1ea"
      },
      "source": [
        "# define input sequence\n",
        "timeseries_data = [110, 125, 133, 146, 158, 172, 187, 196, 210]\n",
        "# choose a number of time steps\n",
        "n_steps = 3\n",
        "# split into samples\n",
        "X, y = prepare_data(timeseries_data, n_steps)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "YQMGJODAk95f",
        "outputId": "5fcb7182-7e69-4f27-a0cb-4884b7bac114"
      },
      "source": [
        "\n",
        "print(X),print(y)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[[110 125 133]\n",
            " [125 133 146]\n",
            " [133 146 158]\n",
            " [146 158 172]\n",
            " [158 172 187]\n",
            " [172 187 196]]\n",
            "[146 158 172 187 196 210]\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(None, None)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "A-FfjX7tlIBU",
        "outputId": "c5fc407c-e289-4b7d-ebe9-73cb792a3a22"
      },
      "source": [
        "X.shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(6, 3)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ejUKHvJ0lPxx"
      },
      "source": [
        "\n",
        "\n",
        "# reshape from [samples, timesteps] into [samples, timesteps, features]\n",
        "n_features = 1\n",
        "X = X.reshape((X.shape[0], X.shape[1], n_features))\n",
        "\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qOrqw4lflawy"
      },
      "source": [
        "\n",
        "Building LSTM Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "q7EIdg6olb23",
        "outputId": "fe4c0e34-2206-4a91-b9fc-ea5ae991b5d5"
      },
      "source": [
        "\n",
        "\n",
        "# define model\n",
        "model = Sequential()\n",
        "model.add(LSTM(50, activation='relu', return_sequences=True, input_shape=(n_steps, n_features)))\n",
        "model.add(LSTM(50, activation='relu'))\n",
        "model.add(Dense(1))\n",
        "model.compile(optimizer='adam', loss='mse')\n",
        "# fit model\n",
        "model.fit(X, y, epochs=300, verbose=1)\n",
        "\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/300\n",
            "1/1 [==============================] - 2s 2s/step - loss: 33709.1758\n",
            "Epoch 2/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 33397.0664\n",
            "Epoch 3/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 33092.2695\n",
            "Epoch 4/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 32772.0625\n",
            "Epoch 5/300\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 32429.9766\n",
            "Epoch 6/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 32062.2969\n",
            "Epoch 7/300\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 31645.2754\n",
            "Epoch 8/300\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 31145.1855\n",
            "Epoch 9/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 30513.2734\n",
            "Epoch 10/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 29805.2969\n",
            "Epoch 11/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 29099.1777\n",
            "Epoch 12/300\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 28416.2598\n",
            "Epoch 13/300\n",
            "1/1 [==============================] - 0s 12ms/step - loss: 27760.8145\n",
            "Epoch 14/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 27068.9199\n",
            "Epoch 15/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 26301.5137\n",
            "Epoch 16/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 25472.7324\n",
            "Epoch 17/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 24625.2578\n",
            "Epoch 18/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 23802.0469\n",
            "Epoch 19/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 22995.5938\n",
            "Epoch 20/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 22191.2109\n",
            "Epoch 21/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 21364.8574\n",
            "Epoch 22/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 20472.8262\n",
            "Epoch 23/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 19475.1230\n",
            "Epoch 24/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 18311.0938\n",
            "Epoch 25/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 16846.5781\n",
            "Epoch 26/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 14913.0264\n",
            "Epoch 27/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 12935.8857\n",
            "Epoch 28/300\n",
            "1/1 [==============================] - 0s 15ms/step - loss: 11282.6064\n",
            "Epoch 29/300\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 9818.3115\n",
            "Epoch 30/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 8426.0137\n",
            "Epoch 31/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 7052.7168\n",
            "Epoch 32/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 5658.7856\n",
            "Epoch 33/300\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 4072.4387\n",
            "Epoch 34/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 2584.2229\n",
            "Epoch 35/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 1404.5238\n",
            "Epoch 36/300\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 573.1956\n",
            "Epoch 37/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 119.0676\n",
            "Epoch 38/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 42.8409\n",
            "Epoch 39/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 259.1221\n",
            "Epoch 40/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 577.9531\n",
            "Epoch 41/300\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 853.8751\n",
            "Epoch 42/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 1018.7671\n",
            "Epoch 43/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 1058.0526\n",
            "Epoch 44/300\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 973.6184\n",
            "Epoch 45/300\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 750.1340\n",
            "Epoch 46/300\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 482.0172\n",
            "Epoch 47/300\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 302.2688\n",
            "Epoch 48/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 183.8743\n",
            "Epoch 49/300\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 105.9019\n",
            "Epoch 50/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 54.9588\n",
            "Epoch 51/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 26.7504\n",
            "Epoch 52/300\n",
            "1/1 [==============================] - 0s 16ms/step - loss: 18.8121\n",
            "Epoch 53/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 26.8424\n",
            "Epoch 54/300\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 44.9539\n",
            "Epoch 55/300\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 66.8427\n",
            "Epoch 56/300\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 86.9186\n",
            "Epoch 57/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 101.1292\n",
            "Epoch 58/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 107.3343\n",
            "Epoch 59/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 105.2368\n",
            "Epoch 60/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 96.0196\n",
            "Epoch 61/300\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 81.8396\n",
            "Epoch 62/300\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 65.3060\n",
            "Epoch 63/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 48.9833\n",
            "Epoch 64/300\n",
            "1/1 [==============================] - 0s 12ms/step - loss: 35.0601\n",
            "Epoch 65/300\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 25.0375\n",
            "Epoch 66/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 19.5699\n",
            "Epoch 67/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 18.4742\n",
            "Epoch 68/300\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 20.8463\n",
            "Epoch 69/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 25.2725\n",
            "Epoch 70/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 30.2308\n",
            "Epoch 71/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 34.4240\n",
            "Epoch 72/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 36.9727\n",
            "Epoch 73/300\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 37.4934\n",
            "Epoch 74/300\n",
            "1/1 [==============================] - 0s 18ms/step - loss: 36.0745\n",
            "Epoch 75/300\n",
            "1/1 [==============================] - 0s 27ms/step - loss: 33.1673\n",
            "Epoch 76/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 29.4330\n",
            "Epoch 77/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 25.5800\n",
            "Epoch 78/300\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 22.2232\n",
            "Epoch 79/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 19.7879\n",
            "Epoch 80/300\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 18.4551\n",
            "Epoch 81/300\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 18.1794\n",
            "Epoch 82/300\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 18.7321\n",
            "Epoch 83/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 19.7762\n",
            "Epoch 84/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 20.9495\n",
            "Epoch 85/300\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 21.9365\n",
            "Epoch 86/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 22.5199\n",
            "Epoch 87/300\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 22.6040\n",
            "Epoch 88/300\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 22.2111\n",
            "Epoch 89/300\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 21.4567\n",
            "Epoch 90/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 20.5111\n",
            "Epoch 91/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 19.5559\n",
            "Epoch 92/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 18.7479\n",
            "Epoch 93/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 18.1901\n",
            "Epoch 94/300\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 17.9211\n",
            "Epoch 95/300\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 17.9167\n",
            "Epoch 96/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 18.1050\n",
            "Epoch 97/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 18.3887\n",
            "Epoch 98/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 18.6694\n",
            "Epoch 99/300\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 18.8681\n",
            "Epoch 100/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 18.9381\n",
            "Epoch 101/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 18.8698\n",
            "Epoch 102/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 18.6860\n",
            "Epoch 103/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 18.4316\n",
            "Epoch 104/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 18.1595\n",
            "Epoch 105/300\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 17.9184\n",
            "Epoch 106/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 17.7421\n",
            "Epoch 107/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 17.6452\n",
            "Epoch 108/300\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 17.6231\n",
            "Epoch 109/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 17.6564\n",
            "Epoch 110/300\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 17.7172\n",
            "Epoch 111/300\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 17.7773\n",
            "Epoch 112/300\n",
            "1/1 [==============================] - 0s 15ms/step - loss: 17.8142\n",
            "Epoch 113/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 17.8147\n",
            "Epoch 114/300\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 17.7767\n",
            "Epoch 115/300\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 17.7072\n",
            "Epoch 116/300\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 17.6198\n",
            "Epoch 117/300\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 17.5295\n",
            "Epoch 118/300\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 17.4500\n",
            "Epoch 119/300\n",
            "1/1 [==============================] - 0s 12ms/step - loss: 17.3902\n",
            "Epoch 120/300\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 17.3531\n",
            "Epoch 121/300\n",
            "1/1 [==============================] - 0s 17ms/step - loss: 17.3359\n",
            "Epoch 122/300\n",
            "1/1 [==============================] - 0s 14ms/step - loss: 17.3323\n",
            "Epoch 123/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 17.3335\n",
            "Epoch 124/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 17.3316\n",
            "Epoch 125/300\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 17.3210\n",
            "Epoch 126/300\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 17.2989\n",
            "Epoch 127/300\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 17.2660\n",
            "Epoch 128/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 17.2256\n",
            "Epoch 129/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 17.1822\n",
            "Epoch 130/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 17.1402\n",
            "Epoch 131/300\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 17.1031\n",
            "Epoch 132/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 17.0723\n",
            "Epoch 133/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 17.0478\n",
            "Epoch 134/300\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 17.0278\n",
            "Epoch 135/300\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 17.0098\n",
            "Epoch 136/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 16.9913\n",
            "Epoch 137/300\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 16.9703\n",
            "Epoch 138/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 16.9460\n",
            "Epoch 139/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 16.9181\n",
            "Epoch 140/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 16.8878\n",
            "Epoch 141/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 16.8563\n",
            "Epoch 142/300\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 16.8249\n",
            "Epoch 143/300\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 16.7950\n",
            "Epoch 144/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 16.7668\n",
            "Epoch 145/300\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 16.7403\n",
            "Epoch 146/300\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 16.7151\n",
            "Epoch 147/300\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 16.6903\n",
            "Epoch 148/300\n",
            "1/1 [==============================] - 0s 14ms/step - loss: 16.6653\n",
            "Epoch 149/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 16.6392\n",
            "Epoch 150/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 16.6119\n",
            "Epoch 151/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 16.5834\n",
            "Epoch 152/300\n",
            "1/1 [==============================] - 0s 14ms/step - loss: 16.5542\n",
            "Epoch 153/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 16.5243\n",
            "Epoch 154/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 16.4946\n",
            "Epoch 155/300\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 16.4650\n",
            "Epoch 156/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 16.4358\n",
            "Epoch 157/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 16.4070\n",
            "Epoch 158/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 16.3784\n",
            "Epoch 159/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 16.3497\n",
            "Epoch 160/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 16.3206\n",
            "Epoch 161/300\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 16.2910\n",
            "Epoch 162/300\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 16.2608\n",
            "Epoch 163/300\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 16.2301\n",
            "Epoch 164/300\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 16.1991\n",
            "Epoch 165/300\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 16.1678\n",
            "Epoch 166/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 16.1363\n",
            "Epoch 167/300\n",
            "1/1 [==============================] - 0s 14ms/step - loss: 16.1049\n",
            "Epoch 168/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 16.0733\n",
            "Epoch 169/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 16.0417\n",
            "Epoch 170/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 16.0098\n",
            "Epoch 171/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 15.9775\n",
            "Epoch 172/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 15.9450\n",
            "Epoch 173/300\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 15.9121\n",
            "Epoch 174/300\n",
            "1/1 [==============================] - 0s 14ms/step - loss: 15.8787\n",
            "Epoch 175/300\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 15.8449\n",
            "Epoch 176/300\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 15.8108\n",
            "Epoch 177/300\n",
            "1/1 [==============================] - 0s 23ms/step - loss: 15.7762\n",
            "Epoch 178/300\n",
            "1/1 [==============================] - 0s 18ms/step - loss: 15.7414\n",
            "Epoch 179/300\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 15.7062\n",
            "Epoch 180/300\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 15.6706\n",
            "Epoch 181/300\n",
            "1/1 [==============================] - 0s 25ms/step - loss: 15.6346\n",
            "Epoch 182/300\n",
            "1/1 [==============================] - 0s 27ms/step - loss: 15.5978\n",
            "Epoch 183/300\n",
            "1/1 [==============================] - 0s 19ms/step - loss: 15.5604\n",
            "Epoch 184/300\n",
            "1/1 [==============================] - 0s 16ms/step - loss: 15.5219\n",
            "Epoch 185/300\n",
            "1/1 [==============================] - 0s 30ms/step - loss: 15.4823\n",
            "Epoch 186/300\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 15.4408\n",
            "Epoch 187/300\n",
            "1/1 [==============================] - 0s 22ms/step - loss: 15.3966\n",
            "Epoch 188/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 15.3476\n",
            "Epoch 189/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 15.2901\n",
            "Epoch 190/300\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 15.2156\n",
            "Epoch 191/300\n",
            "1/1 [==============================] - 0s 13ms/step - loss: 15.1057\n",
            "Epoch 192/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 14.9224\n",
            "Epoch 193/300\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 14.6191\n",
            "Epoch 194/300\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 14.3760\n",
            "Epoch 195/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 14.7425\n",
            "Epoch 196/300\n",
            "1/1 [==============================] - 0s 14ms/step - loss: 14.1291\n",
            "Epoch 197/300\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 14.1555\n",
            "Epoch 198/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 14.2790\n",
            "Epoch 199/300\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 14.2685\n",
            "Epoch 200/300\n",
            "1/1 [==============================] - 0s 13ms/step - loss: 14.0593\n",
            "Epoch 201/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 13.7166\n",
            "Epoch 202/300\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 13.7968\n",
            "Epoch 203/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 13.8445\n",
            "Epoch 204/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 13.4842\n",
            "Epoch 205/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 13.5666\n",
            "Epoch 206/300\n",
            "1/1 [==============================] - 0s 20ms/step - loss: 13.6044\n",
            "Epoch 207/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 13.3730\n",
            "Epoch 208/300\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 13.1130\n",
            "Epoch 209/300\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 13.9179\n",
            "Epoch 210/300\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 12.9419\n",
            "Epoch 211/300\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 13.1478\n",
            "Epoch 212/300\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 13.2459\n",
            "Epoch 213/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 13.0201\n",
            "Epoch 214/300\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 12.8191\n",
            "Epoch 215/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 12.9827\n",
            "Epoch 216/300\n",
            "1/1 [==============================] - 0s 12ms/step - loss: 12.8372\n",
            "Epoch 217/300\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 12.6116\n",
            "Epoch 218/300\n",
            "1/1 [==============================] - 0s 14ms/step - loss: 12.6466\n",
            "Epoch 219/300\n",
            "1/1 [==============================] - 0s 14ms/step - loss: 12.6222\n",
            "Epoch 220/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 12.4456\n",
            "Epoch 221/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 12.3005\n",
            "Epoch 222/300\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 12.3135\n",
            "Epoch 223/300\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 12.2200\n",
            "Epoch 224/300\n",
            "1/1 [==============================] - 0s 13ms/step - loss: 12.0474\n",
            "Epoch 225/300\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 11.9913\n",
            "Epoch 226/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 11.9404\n",
            "Epoch 227/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 11.8085\n",
            "Epoch 228/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 11.6682\n",
            "Epoch 229/300\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 11.6036\n",
            "Epoch 230/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 11.5114\n",
            "Epoch 231/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 11.3553\n",
            "Epoch 232/300\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 11.2486\n",
            "Epoch 233/300\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 11.1564\n",
            "Epoch 234/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 11.0149\n",
            "Epoch 235/300\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 10.8694\n",
            "Epoch 236/300\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 10.7640\n",
            "Epoch 237/300\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 10.6266\n",
            "Epoch 238/300\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 10.4642\n",
            "Epoch 239/300\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 10.3337\n",
            "Epoch 240/300\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 10.1899\n",
            "Epoch 241/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 10.0186\n",
            "Epoch 242/300\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 9.8670\n",
            "Epoch 243/300\n",
            "1/1 [==============================] - 0s 13ms/step - loss: 9.7114\n",
            "Epoch 244/300\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 9.5258\n",
            "Epoch 245/300\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 9.3557\n",
            "Epoch 246/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 9.1856\n",
            "Epoch 247/300\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 8.9983\n",
            "Epoch 248/300\n",
            "1/1 [==============================] - 0s 16ms/step - loss: 8.8237\n",
            "Epoch 249/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 8.6481\n",
            "Epoch 250/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 8.4557\n",
            "Epoch 251/300\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 8.2747\n",
            "Epoch 252/300\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 8.0932\n",
            "Epoch 253/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 7.9067\n",
            "Epoch 254/300\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 7.7330\n",
            "Epoch 255/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 7.5553\n",
            "Epoch 256/300\n",
            "1/1 [==============================] - 0s 20ms/step - loss: 7.3756\n",
            "Epoch 257/300\n",
            "1/1 [==============================] - 0s 34ms/step - loss: 7.2072\n",
            "Epoch 258/300\n",
            "1/1 [==============================] - 0s 28ms/step - loss: 7.0377\n",
            "Epoch 259/300\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 6.8747\n",
            "Epoch 260/300\n",
            "1/1 [==============================] - 0s 12ms/step - loss: 6.7193\n",
            "Epoch 261/300\n",
            "1/1 [==============================] - 0s 19ms/step - loss: 6.5620\n",
            "Epoch 262/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 6.4166\n",
            "Epoch 263/300\n",
            "1/1 [==============================] - 0s 15ms/step - loss: 6.2789\n",
            "Epoch 264/300\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 6.1502\n",
            "Epoch 265/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 6.0337\n",
            "Epoch 266/300\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 5.9196\n",
            "Epoch 267/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 5.8095\n",
            "Epoch 268/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 5.7022\n",
            "Epoch 269/300\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 5.5996\n",
            "Epoch 270/300\n",
            "1/1 [==============================] - 0s 14ms/step - loss: 5.5060\n",
            "Epoch 271/300\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 5.4157\n",
            "Epoch 272/300\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 5.3340\n",
            "Epoch 273/300\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 5.2558\n",
            "Epoch 274/300\n",
            "1/1 [==============================] - 0s 20ms/step - loss: 5.1829\n",
            "Epoch 275/300\n",
            "1/1 [==============================] - 0s 16ms/step - loss: 5.1109\n",
            "Epoch 276/300\n",
            "1/1 [==============================] - 0s 21ms/step - loss: 5.0406\n",
            "Epoch 277/300\n",
            "1/1 [==============================] - 0s 18ms/step - loss: 4.9676\n",
            "Epoch 278/300\n",
            "1/1 [==============================] - 0s 14ms/step - loss: 4.8974\n",
            "Epoch 279/300\n",
            "1/1 [==============================] - 0s 17ms/step - loss: 4.8259\n",
            "Epoch 280/300\n",
            "1/1 [==============================] - 0s 13ms/step - loss: 4.7569\n",
            "Epoch 281/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 4.6878\n",
            "Epoch 282/300\n",
            "1/1 [==============================] - 0s 23ms/step - loss: 4.6220\n",
            "Epoch 283/300\n",
            "1/1 [==============================] - 0s 23ms/step - loss: 4.5573\n",
            "Epoch 284/300\n",
            "1/1 [==============================] - 0s 16ms/step - loss: 4.4965\n",
            "Epoch 285/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 4.4386\n",
            "Epoch 286/300\n",
            "1/1 [==============================] - 0s 12ms/step - loss: 4.3848\n",
            "Epoch 287/300\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 4.3333\n",
            "Epoch 288/300\n",
            "1/1 [==============================] - 0s 13ms/step - loss: 4.2848\n",
            "Epoch 289/300\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 4.2377\n",
            "Epoch 290/300\n",
            "1/1 [==============================] - 0s 14ms/step - loss: 4.1914\n",
            "Epoch 291/300\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 4.1456\n",
            "Epoch 292/300\n",
            "1/1 [==============================] - 0s 14ms/step - loss: 4.0997\n",
            "Epoch 293/300\n",
            "1/1 [==============================] - 0s 12ms/step - loss: 4.0547\n",
            "Epoch 294/300\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 4.0101\n",
            "Epoch 295/300\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 3.9669\n",
            "Epoch 296/300\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 3.9248\n",
            "Epoch 297/300\n",
            "1/1 [==============================] - 0s 13ms/step - loss: 3.8843\n",
            "Epoch 298/300\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 3.8448\n",
            "Epoch 299/300\n",
            "1/1 [==============================] - 0s 17ms/step - loss: 3.8065\n",
            "Epoch 300/300\n",
            "1/1 [==============================] - 0s 14ms/step - loss: 3.7691\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tensorflow.python.keras.callbacks.History at 0x7f4a656d0050>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 7
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0UnsNR9dl8vs"
      },
      "source": [
        "\n",
        "Predicting For the next 10 data\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "bw1j56fHl_LP",
        "outputId": "d1e2c448-62e2-4a53-c72f-d288d575988d"
      },
      "source": [
        "# demonstrate prediction for next 10 days\n",
        "x_input = np.array([187, 196, 210])\n",
        "temp_input=list(x_input)\n",
        "lst_output=[]\n",
        "i=0\n",
        "while(i<10):\n",
        "    \n",
        "    if(len(temp_input)>3):\n",
        "        x_input= np.array(temp_input[1:])\n",
        "        print(\"{} day input {}\".format(i,x_input))\n",
        "        #print(x_input)\n",
        "        x_input = x_input.reshape((1, n_steps, n_features))\n",
        "        #print(x_input)\n",
        "        yhat = model.predict(x_input, verbose=0)\n",
        "        print(\"{} day output {}\".format(i,yhat))\n",
        "        temp_input.append(yhat[0][0])\n",
        "        temp_input=temp_input[1:]\n",
        "        #print(temp_input)\n",
        "        lst_output.append(yhat[0][0])\n",
        "        i=i+1\n",
        "    else:\n",
        "        x_input = x_input.reshape((1, n_steps, n_features))\n",
        "        yhat = model.predict(x_input, verbose=0)\n",
        "        print(yhat[0])\n",
        "        temp_input.append(yhat[0][0])\n",
        "        lst_output.append(yhat[0][0])\n",
        "        i=i+1\n",
        "    \n",
        "\n",
        "print(lst_output)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[226.53856]\n",
            "1 day input [196.         210.         226.53855896]\n",
            "1 day output [[236.51779]]\n",
            "2 day input [210.         226.53855896 236.51779175]\n",
            "2 day output [[250.95737]]\n",
            "3 day input [226.53856 236.51779 250.95737]\n",
            "3 day output [[267.21735]]\n",
            "4 day input [236.51779 250.95737 267.21735]\n",
            "4 day output [[279.73462]]\n",
            "5 day input [250.95737 267.21735 279.73462]\n",
            "5 day output [[295.6571]]\n",
            "6 day input [267.21735 279.73462 295.6571 ]\n",
            "6 day output [[312.46304]]\n",
            "7 day input [279.73462 295.6571  312.46304]\n",
            "7 day output [[327.891]]\n",
            "8 day input [295.6571  312.46304 327.891  ]\n",
            "8 day output [[345.77017]]\n",
            "9 day input [312.46304 327.891   345.77017]\n",
            "9 day output [[364.24387]]\n",
            "[226.53856, 236.51779, 250.95737, 267.21735, 279.73462, 295.6571, 312.46304, 327.891, 345.77017, 364.24387]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ivlhGYtAnB6S",
        "outputId": "23f50c68-630a-4d0d-af23-801d47bc2af5"
      },
      "source": [
        "timeseries_data"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[110, 125, 133, 146, 158, 172, 187, 196, 210]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 11
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VqcSwuWsnJRq",
        "outputId": "62256f17-a91f-4049-987c-32383b8e1f66"
      },
      "source": [
        "len(timeseries_data)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "9"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "p3WVfBWUnQTu",
        "outputId": "1cd23fba-1429-4963-a7b1-ff34db948ad4"
      },
      "source": [
        "lst_output\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[226.53856,\n",
              " 236.51779,\n",
              " 250.95737,\n",
              " 267.21735,\n",
              " 279.73462,\n",
              " 295.6571,\n",
              " 312.46304,\n",
              " 327.891,\n",
              " 345.77017,\n",
              " 364.24387]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 13
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WXUxZDsDpwqT"
      },
      "source": [
        "\n",
        "Visualizaing The Output"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7h8zxWnPpdgy"
      },
      "source": [
        "import matplotlib.pyplot as plt"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Bmffv1i3plQ7"
      },
      "source": [
        "day_new=np.arange(1,10)\n",
        "day_pred=np.arange(10,20)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "id": "mgqyuM_1prji",
        "outputId": "8dd3f0ab-c1fc-45c5-e4c0-d1d874aaf218"
      },
      "source": [
        "plt.plot(day_new,timeseries_data)\n",
        "plt.plot(day_pred,lst_output)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<matplotlib.lines.Line2D at 0x7f4a5a3aea90>]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 22
        },
        {
          "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"
          }
        }
      ]
    }
  ]
}