5_TimeSeries_ Sunspot.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "executionInfo": {
     "elapsed": 1004,
     "status": "ok",
     "timestamp": 1609039730789,
     "user": {
      "displayName": "Ly Nguyen",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14Gj1nBUdqgyGPrTg8lXGC8cHEEjZ06cg4eNIe54J=s64",
      "userId": "01585632328413993117"
     },
     "user_tz": -420
    },
    "id": "yZgQj9lHyk-G"
   },
   "outputs": [],
   "source": [
    "import tensorflow as tf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "executionInfo": {
     "elapsed": 1069,
     "status": "ok",
     "timestamp": 1609039732335,
     "user": {
      "displayName": "Ly Nguyen",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14Gj1nBUdqgyGPrTg8lXGC8cHEEjZ06cg4eNIe54J=s64",
      "userId": "01585632328413993117"
     },
     "user_tz": -420
    },
    "id": "odcdI4hFKGHe"
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "def plot_series(time, series, format=\"-\", start=0, end=None):\n",
    "    plt.plot(time[start:end], series[start:end], format)\n",
    "    plt.xlabel(\"Time\")\n",
    "    plt.ylabel(\"Value\")\n",
    "    plt.grid(True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 888,
     "status": "ok",
     "timestamp": 1609039734029,
     "user": {
      "displayName": "Ly Nguyen",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14Gj1nBUdqgyGPrTg8lXGC8cHEEjZ06cg4eNIe54J=s64",
      "userId": "01585632328413993117"
     },
     "user_tz": -420
    },
    "id": "v7InVhV5KKE4",
    "outputId": "ce64b791-276f-4781-c3b6-7b372de6ccaa"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "--2020-12-27 03:28:53--  https://raw.githubusercontent.com/jbrownlee/Datasets/master/daily-min-temperatures.csv\n",
      "Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 151.101.0.133, 151.101.64.133, 151.101.128.133, ...\n",
      "Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|151.101.0.133|:443... connected.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 67921 (66K) [text/plain]\n",
      "Saving to: ‘/tmp/daily-min-temperatures.csv’\n",
      "\n",
      "/tmp/daily-min-temp 100%[===================>]  66.33K  --.-KB/s    in 0.002s  \n",
      "\n",
      "2020-12-27 03:28:53 (38.4 MB/s) - ‘/tmp/daily-min-temperatures.csv’ saved [67921/67921]\n",
      "\n"
     ]
    }
   ],
   "source": [
    "!wget --no-check-certificate \\\n",
    "    https://raw.githubusercontent.com/jbrownlee/Datasets/master/daily-min-temperatures.csv \\\n",
    "    -O /tmp/daily-min-temperatures.csv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 387
    },
    "executionInfo": {
     "elapsed": 1159,
     "status": "ok",
     "timestamp": 1609039740166,
     "user": {
      "displayName": "Ly Nguyen",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14Gj1nBUdqgyGPrTg8lXGC8cHEEjZ06cg4eNIe54J=s64",
      "userId": "01585632328413993117"
     },
     "user_tz": -420
    },
    "id": "vRo-JVJGKMuF",
    "outputId": "4babbaa2-56d0-43e2-960b-e9e154009db6"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import csv\n",
    "time_step = []\n",
    "temps = []\n",
    "\n",
    "with open('/content/Sunspots.csv') as csvfile:\n",
    "  reader = csv.reader(csvfile, delimiter=',')\n",
    "  next(reader)\n",
    "  step=0\n",
    "  for row in reader:\n",
    "    temps.append(float(row[2]))\n",
    "    time_step.append(step)\n",
    "    step = step + 1\n",
    "\n",
    "series = np.array(temps)\n",
    "time = np.array(time_step)\n",
    "plt.figure(figsize=(10, 6))\n",
    "plot_series(time, series)\n",
    "\n",
    "min = np.min(series)\n",
    "max = np.max(series)\n",
    "series -= min\n",
    "series /= max"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {
    "executionInfo": {
     "elapsed": 877,
     "status": "ok",
     "timestamp": 1609040168311,
     "user": {
      "displayName": "Ly Nguyen",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14Gj1nBUdqgyGPrTg8lXGC8cHEEjZ06cg4eNIe54J=s64",
      "userId": "01585632328413993117"
     },
     "user_tz": -420
    },
    "id": "K6oK_iFHKO82"
   },
   "outputs": [],
   "source": [
    "split_time = 3000\n",
    "time_train = time[:split_time]\n",
    "x_train = series[:split_time]\n",
    "time_valid = time[split_time:]\n",
    "x_valid = series[split_time:]\n",
    "\n",
    "window_size = 30\n",
    "batch_size = 32\n",
    "shuffle_buffer_size = 1000"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "executionInfo": {
     "elapsed": 828,
     "status": "ok",
     "timestamp": 1609040170313,
     "user": {
      "displayName": "Ly Nguyen",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14Gj1nBUdqgyGPrTg8lXGC8cHEEjZ06cg4eNIe54J=s64",
      "userId": "01585632328413993117"
     },
     "user_tz": -420
    },
    "id": "oJ59iLswKStm"
   },
   "outputs": [],
   "source": [
    "def windowed_dataset(series, window_size, batch_size, shuffle_buffer):\n",
    "    series = tf.expand_dims(series, axis=-1)\n",
    "    ds = tf.data.Dataset.from_tensor_slices(series)\n",
    "    ds = ds.window(window_size + 1, shift=1, drop_remainder=True)\n",
    "    ds = ds.flat_map(lambda w: w.batch(window_size + 1))\n",
    "    ds = ds.shuffle(shuffle_buffer)\n",
    "    ds = ds.map(lambda w: (w[:-1], w[1:]))\n",
    "    return ds.batch(batch_size).prefetch(1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "executionInfo": {
     "elapsed": 890,
     "status": "ok",
     "timestamp": 1609039751579,
     "user": {
      "displayName": "Ly Nguyen",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14Gj1nBUdqgyGPrTg8lXGC8cHEEjZ06cg4eNIe54J=s64",
      "userId": "01585632328413993117"
     },
     "user_tz": -420
    },
    "id": "IqVxOB2TKTRt"
   },
   "outputs": [],
   "source": [
    "def model_forecast(model, series, window_size):\n",
    "    ds = tf.data.Dataset.from_tensor_slices(series)\n",
    "    ds = ds.window(window_size, shift=1, drop_remainder=True)\n",
    "    ds = ds.flat_map(lambda w: w.batch(window_size))\n",
    "    ds = ds.batch(32).prefetch(1)\n",
    "    forecast = model.predict(ds)\n",
    "    return forecast"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 141178,
     "status": "ok",
     "timestamp": 1609040740607,
     "user": {
      "displayName": "Ly Nguyen",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14Gj1nBUdqgyGPrTg8lXGC8cHEEjZ06cg4eNIe54J=s64",
      "userId": "01585632328413993117"
     },
     "user_tz": -420
    },
    "id": "GJccx_aJKsqv",
    "outputId": "56d7dd22-e3c9-4340-e7f6-03690b0e3954"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/150\n",
      "93/93 [==============================] - 3s 10ms/step - loss: 0.0339 - mae: 0.1989\n",
      "Epoch 2/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0338 - mae: 0.1986\n",
      "Epoch 3/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0333 - mae: 0.1958\n",
      "Epoch 4/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0314 - mae: 0.1890\n",
      "Epoch 5/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0309 - mae: 0.1862\n",
      "Epoch 6/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0302 - mae: 0.1827\n",
      "Epoch 7/150\n",
      "93/93 [==============================] - 1s 10ms/step - loss: 0.0302 - mae: 0.1822\n",
      "Epoch 8/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0288 - mae: 0.1766\n",
      "Epoch 9/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0283 - mae: 0.1752\n",
      "Epoch 10/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0291 - mae: 0.1769\n",
      "Epoch 11/150\n",
      "93/93 [==============================] - 1s 8ms/step - loss: 0.0263 - mae: 0.1667\n",
      "Epoch 12/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0264 - mae: 0.1648\n",
      "Epoch 13/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0261 - mae: 0.1654\n",
      "Epoch 14/150\n",
      "93/93 [==============================] - 1s 8ms/step - loss: 0.0244 - mae: 0.1588\n",
      "Epoch 15/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0246 - mae: 0.1598\n",
      "Epoch 16/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0238 - mae: 0.1566\n",
      "Epoch 17/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0238 - mae: 0.1559\n",
      "Epoch 18/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0230 - mae: 0.1539\n",
      "Epoch 19/150\n",
      "93/93 [==============================] - 1s 8ms/step - loss: 0.0226 - mae: 0.1527\n",
      "Epoch 20/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0220 - mae: 0.1493\n",
      "Epoch 21/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0217 - mae: 0.1492\n",
      "Epoch 22/150\n",
      "93/93 [==============================] - 1s 8ms/step - loss: 0.0207 - mae: 0.1448\n",
      "Epoch 23/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0210 - mae: 0.1472\n",
      "Epoch 24/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0204 - mae: 0.1437\n",
      "Epoch 25/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0196 - mae: 0.1422\n",
      "Epoch 26/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0185 - mae: 0.1375\n",
      "Epoch 27/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0184 - mae: 0.1370\n",
      "Epoch 28/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0179 - mae: 0.1348\n",
      "Epoch 29/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0177 - mae: 0.1352\n",
      "Epoch 30/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0174 - mae: 0.1337\n",
      "Epoch 31/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0187 - mae: 0.1397\n",
      "Epoch 32/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0174 - mae: 0.1343\n",
      "Epoch 33/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0170 - mae: 0.1330\n",
      "Epoch 34/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0170 - mae: 0.1337\n",
      "Epoch 35/150\n",
      "93/93 [==============================] - 1s 8ms/step - loss: 0.0169 - mae: 0.1328\n",
      "Epoch 36/150\n",
      "93/93 [==============================] - 1s 8ms/step - loss: 0.0170 - mae: 0.1333\n",
      "Epoch 37/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0160 - mae: 0.1300\n",
      "Epoch 38/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0160 - mae: 0.1302\n",
      "Epoch 39/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0162 - mae: 0.1307\n",
      "Epoch 40/150\n",
      "93/93 [==============================] - 1s 10ms/step - loss: 0.0158 - mae: 0.1298\n",
      "Epoch 41/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0153 - mae: 0.1281\n",
      "Epoch 42/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0154 - mae: 0.1290\n",
      "Epoch 43/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0152 - mae: 0.1279\n",
      "Epoch 44/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0156 - mae: 0.1302\n",
      "Epoch 45/150\n",
      "93/93 [==============================] - 1s 10ms/step - loss: 0.0151 - mae: 0.1279\n",
      "Epoch 46/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0151 - mae: 0.1285\n",
      "Epoch 47/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0150 - mae: 0.1283\n",
      "Epoch 48/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0147 - mae: 0.1272\n",
      "Epoch 49/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0145 - mae: 0.1270\n",
      "Epoch 50/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0146 - mae: 0.1267\n",
      "Epoch 51/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0145 - mae: 0.1273\n",
      "Epoch 52/150\n",
      "93/93 [==============================] - 1s 8ms/step - loss: 0.0142 - mae: 0.1258\n",
      "Epoch 53/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0142 - mae: 0.1262\n",
      "Epoch 54/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0141 - mae: 0.1255\n",
      "Epoch 55/150\n",
      "93/93 [==============================] - 1s 8ms/step - loss: 0.0145 - mae: 0.1277\n",
      "Epoch 56/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0144 - mae: 0.1281\n",
      "Epoch 57/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0144 - mae: 0.1275\n",
      "Epoch 58/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0139 - mae: 0.1261\n",
      "Epoch 59/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0143 - mae: 0.1282\n",
      "Epoch 60/150\n",
      "93/93 [==============================] - 1s 8ms/step - loss: 0.0138 - mae: 0.1255\n",
      "Epoch 61/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0139 - mae: 0.1267\n",
      "Epoch 62/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0140 - mae: 0.1270\n",
      "Epoch 63/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0140 - mae: 0.1272\n",
      "Epoch 64/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0134 - mae: 0.1245\n",
      "Epoch 65/150\n",
      "93/93 [==============================] - 1s 10ms/step - loss: 0.0137 - mae: 0.1263\n",
      "Epoch 66/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0141 - mae: 0.1281\n",
      "Epoch 67/150\n",
      "93/93 [==============================] - 1s 8ms/step - loss: 0.0137 - mae: 0.1268\n",
      "Epoch 68/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0134 - mae: 0.1253\n",
      "Epoch 69/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0135 - mae: 0.1256\n",
      "Epoch 70/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0135 - mae: 0.1261\n",
      "Epoch 71/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0133 - mae: 0.1252\n",
      "Epoch 72/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0136 - mae: 0.1268\n",
      "Epoch 73/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0133 - mae: 0.1254\n",
      "Epoch 74/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0131 - mae: 0.1249\n",
      "Epoch 75/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0128 - mae: 0.1236\n",
      "Epoch 76/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0136 - mae: 0.1272\n",
      "Epoch 77/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0132 - mae: 0.1251\n",
      "Epoch 78/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0130 - mae: 0.1250\n",
      "Epoch 79/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0130 - mae: 0.1253\n",
      "Epoch 80/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0132 - mae: 0.1261\n",
      "Epoch 81/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0130 - mae: 0.1254\n",
      "Epoch 82/150\n",
      "93/93 [==============================] - 1s 10ms/step - loss: 0.0131 - mae: 0.1257\n",
      "Epoch 83/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0131 - mae: 0.1257\n",
      "Epoch 84/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0130 - mae: 0.1255\n",
      "Epoch 85/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0126 - mae: 0.1235\n",
      "Epoch 86/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0132 - mae: 0.1268\n",
      "Epoch 87/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0133 - mae: 0.1269\n",
      "Epoch 88/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0128 - mae: 0.1251\n",
      "Epoch 89/150\n",
      "93/93 [==============================] - 1s 8ms/step - loss: 0.0132 - mae: 0.1266\n",
      "Epoch 90/150\n",
      "93/93 [==============================] - 1s 8ms/step - loss: 0.0128 - mae: 0.1252\n",
      "Epoch 91/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0127 - mae: 0.1243\n",
      "Epoch 92/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0128 - mae: 0.1254\n",
      "Epoch 93/150\n",
      "93/93 [==============================] - 1s 8ms/step - loss: 0.0132 - mae: 0.1268\n",
      "Epoch 94/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0130 - mae: 0.1262\n",
      "Epoch 95/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0128 - mae: 0.1258\n",
      "Epoch 96/150\n",
      "93/93 [==============================] - 1s 10ms/step - loss: 0.0129 - mae: 0.1262\n",
      "Epoch 97/150\n",
      "93/93 [==============================] - 1s 8ms/step - loss: 0.0127 - mae: 0.1249\n",
      "Epoch 98/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0128 - mae: 0.1261\n",
      "Epoch 99/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0128 - mae: 0.1259\n",
      "Epoch 100/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0128 - mae: 0.1261\n",
      "Epoch 101/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0132 - mae: 0.1274\n",
      "Epoch 102/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0126 - mae: 0.1252\n",
      "Epoch 103/150\n",
      "93/93 [==============================] - 1s 8ms/step - loss: 0.0130 - mae: 0.1266\n",
      "Epoch 104/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0126 - mae: 0.1248\n",
      "Epoch 105/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0125 - mae: 0.1245\n",
      "Epoch 106/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0130 - mae: 0.1271\n",
      "Epoch 107/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0128 - mae: 0.1264\n",
      "Epoch 108/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0128 - mae: 0.1265\n",
      "Epoch 109/150\n",
      "93/93 [==============================] - 1s 8ms/step - loss: 0.0128 - mae: 0.1263\n",
      "Epoch 110/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0129 - mae: 0.1267\n",
      "Epoch 111/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0132 - mae: 0.1284\n",
      "Epoch 112/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0127 - mae: 0.1259\n",
      "Epoch 113/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0126 - mae: 0.1252\n",
      "Epoch 114/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0129 - mae: 0.1271\n",
      "Epoch 115/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0128 - mae: 0.1267\n",
      "Epoch 116/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0128 - mae: 0.1263\n",
      "Epoch 117/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0128 - mae: 0.1262\n",
      "Epoch 118/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0126 - mae: 0.1259\n",
      "Epoch 119/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0128 - mae: 0.1270\n",
      "Epoch 120/150\n",
      "93/93 [==============================] - 1s 8ms/step - loss: 0.0129 - mae: 0.1275\n",
      "Epoch 121/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0129 - mae: 0.1273\n",
      "Epoch 122/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0127 - mae: 0.1272\n",
      "Epoch 123/150\n",
      "93/93 [==============================] - 1s 10ms/step - loss: 0.0130 - mae: 0.1278\n",
      "Epoch 124/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0129 - mae: 0.1277\n",
      "Epoch 125/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0128 - mae: 0.1271\n",
      "Epoch 126/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0127 - mae: 0.1268\n",
      "Epoch 127/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0133 - mae: 0.1297\n",
      "Epoch 128/150\n",
      "93/93 [==============================] - 1s 10ms/step - loss: 0.0129 - mae: 0.1276\n",
      "Epoch 129/150\n",
      "93/93 [==============================] - 1s 10ms/step - loss: 0.0127 - mae: 0.1270\n",
      "Epoch 130/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0130 - mae: 0.1284\n",
      "Epoch 131/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0126 - mae: 0.1263\n",
      "Epoch 132/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0129 - mae: 0.1270\n",
      "Epoch 133/150\n",
      "93/93 [==============================] - 1s 10ms/step - loss: 0.0124 - mae: 0.1261\n",
      "Epoch 134/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0131 - mae: 0.1290\n",
      "Epoch 135/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0129 - mae: 0.1277\n",
      "Epoch 136/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0125 - mae: 0.1259\n",
      "Epoch 137/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0127 - mae: 0.1276\n",
      "Epoch 138/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0128 - mae: 0.1273\n",
      "Epoch 139/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0129 - mae: 0.1279\n",
      "Epoch 140/150\n",
      "93/93 [==============================] - 1s 8ms/step - loss: 0.0125 - mae: 0.1262\n",
      "Epoch 141/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0130 - mae: 0.1284\n",
      "Epoch 142/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0128 - mae: 0.1273\n",
      "Epoch 143/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0124 - mae: 0.1261\n",
      "Epoch 144/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0127 - mae: 0.1272\n",
      "Epoch 145/150\n",
      "93/93 [==============================] - 1s 8ms/step - loss: 0.0124 - mae: 0.1258\n",
      "Epoch 146/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0125 - mae: 0.1264\n",
      "Epoch 147/150\n",
      "93/93 [==============================] - 1s 10ms/step - loss: 0.0126 - mae: 0.1267\n",
      "Epoch 148/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0126 - mae: 0.1265\n",
      "Epoch 149/150\n",
      "93/93 [==============================] - 1s 10ms/step - loss: 0.0127 - mae: 0.1273\n",
      "Epoch 150/150\n",
      "93/93 [==============================] - 1s 9ms/step - loss: 0.0125 - mae: 0.1263\n"
     ]
    }
   ],
   "source": [
    "\n",
    "tf.keras.backend.clear_session()\n",
    "tf.random.set_seed(51)\n",
    "np.random.seed(51)\n",
    "\n",
    "train_set = windowed_dataset(x_train, window_size=window_size, batch_size=batch_size, shuffle_buffer=shuffle_buffer_size)\n",
    "\n",
    "# train_set = windowed_dataset(x_train, window_size=60, batch_size=100, shuffle_buffer=shuffle_buffer_size)\n",
    "model = tf.keras.models.Sequential([\n",
    "  tf.keras.layers.Conv1D(filters=60, kernel_size=5,\n",
    "                      strides=1, padding=\"causal\",\n",
    "                      activation=\"relu\",\n",
    "                      input_shape=[None, 1]),\n",
    "  tf.keras.layers.LSTM(60, return_sequences=True),\n",
    "  tf.keras.layers.LSTM(60, return_sequences=True),\n",
    "  tf.keras.layers.Dense(30, activation=\"relu\"),\n",
    "  tf.keras.layers.Dense(10, activation=\"relu\"),\n",
    "  tf.keras.layers.Dense(1),\n",
    "])\n",
    "\n",
    "\n",
    "optimizer = tf.keras.optimizers.SGD(lr=1e-5, momentum=0.9)\n",
    "model.compile(loss=tf.keras.losses.Huber(),\n",
    "              optimizer=optimizer,\n",
    "              metrics=[\"mae\"])\n",
    "history = model.fit(train_set,epochs=150)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {
    "executionInfo": {
     "elapsed": 946,
     "status": "ok",
     "timestamp": 1609040756355,
     "user": {
      "displayName": "Ly Nguyen",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14Gj1nBUdqgyGPrTg8lXGC8cHEEjZ06cg4eNIe54J=s64",
      "userId": "01585632328413993117"
     },
     "user_tz": -420
    },
    "id": "iGCduDVPJfXx"
   },
   "outputs": [],
   "source": [
    "model.save(\"mymodel.h5\")"
   ]
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "authorship_tag": "ABX9TyN5sKpuVGGOYAc72AkPGIXU",
   "collapsed_sections": [],
   "name": "5_TimeSeries_ Sunspot.ipynb",
   "provenance": []
  },
  "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"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}