engine_rul_v19.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import tensorflow as tf\n",
    "from sklearn.preprocessing import StandardScaler,MinMaxScaler\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import mean_absolute_error,mean_squared_error\n",
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import LSTM, Dense, TimeDistributed,SimpleRNN,Dropout,GRU\n",
    "from tensorflow.keras.callbacks import EarlyStopping\n",
    "from tensorflow.keras.preprocessing.sequence import TimeseriesGenerator\n",
    "from tensorflow.keras.callbacks import ReduceLROnPlateau\n",
    "from tensorflow.keras.optimizers import Adam"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[PhysicalDevice(name='/physical_device:GPU:0', device_type='GPU')]"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tf.config.list_physical_devices('GPU')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3.9.20 (main, Oct  3 2024, 07:38:01) [MSC v.1929 64 bit (AMD64)]\n",
      "2.8.0\n"
     ]
    }
   ],
   "source": [
    "import sys\n",
    "print (sys.version)\n",
    " # 3.6.4 |Anaconda custom (64-bit)| (default, Jan 16 2018, 10:22:32) [MSC v.1900 64 bit (AMD64)]\n",
    "print(tf.__version__)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_excel(\"D:\\\\Projects\\\\BDA\\\\aircraft_reliability\\\\data\\\\PM_train.xlsx\")\n",
    "df_truth = pd.read_excel(\"D:\\\\Projects\\\\BDA\\\\aircraft_reliability\\\\data\\\\PM_truth.xlsx\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>cycle</th>\n",
       "      <th>setting1</th>\n",
       "      <th>setting2</th>\n",
       "      <th>setting3</th>\n",
       "      <th>s1</th>\n",
       "      <th>s2</th>\n",
       "      <th>s3</th>\n",
       "      <th>s4</th>\n",
       "      <th>s5</th>\n",
       "      <th>...</th>\n",
       "      <th>s13</th>\n",
       "      <th>s14</th>\n",
       "      <th>s15</th>\n",
       "      <th>s16</th>\n",
       "      <th>s17</th>\n",
       "      <th>s18</th>\n",
       "      <th>s19</th>\n",
       "      <th>s20</th>\n",
       "      <th>s21</th>\n",
       "      <th>more</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>-0.0007</td>\n",
       "      <td>-0.0004</td>\n",
       "      <td>100</td>\n",
       "      <td>518.67</td>\n",
       "      <td>641.82</td>\n",
       "      <td>1589.70</td>\n",
       "      <td>1400.60</td>\n",
       "      <td>14.62</td>\n",
       "      <td>...</td>\n",
       "      <td>2388.02</td>\n",
       "      <td>8138.62</td>\n",
       "      <td>8.4195</td>\n",
       "      <td>0.03</td>\n",
       "      <td>392</td>\n",
       "      <td>2388</td>\n",
       "      <td>100</td>\n",
       "      <td>39.06</td>\n",
       "      <td>23.4190</td>\n",
       "      <td>98</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.0019</td>\n",
       "      <td>-0.0003</td>\n",
       "      <td>100</td>\n",
       "      <td>518.67</td>\n",
       "      <td>642.15</td>\n",
       "      <td>1591.82</td>\n",
       "      <td>1403.14</td>\n",
       "      <td>14.62</td>\n",
       "      <td>...</td>\n",
       "      <td>2388.07</td>\n",
       "      <td>8131.49</td>\n",
       "      <td>8.4318</td>\n",
       "      <td>0.03</td>\n",
       "      <td>392</td>\n",
       "      <td>2388</td>\n",
       "      <td>100</td>\n",
       "      <td>39.00</td>\n",
       "      <td>23.4236</td>\n",
       "      <td>98</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>-0.0043</td>\n",
       "      <td>0.0003</td>\n",
       "      <td>100</td>\n",
       "      <td>518.67</td>\n",
       "      <td>642.35</td>\n",
       "      <td>1587.99</td>\n",
       "      <td>1404.20</td>\n",
       "      <td>14.62</td>\n",
       "      <td>...</td>\n",
       "      <td>2388.03</td>\n",
       "      <td>8133.23</td>\n",
       "      <td>8.4178</td>\n",
       "      <td>0.03</td>\n",
       "      <td>390</td>\n",
       "      <td>2388</td>\n",
       "      <td>100</td>\n",
       "      <td>38.95</td>\n",
       "      <td>23.3442</td>\n",
       "      <td>98</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "      <td>0.0007</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>100</td>\n",
       "      <td>518.67</td>\n",
       "      <td>642.35</td>\n",
       "      <td>1582.79</td>\n",
       "      <td>1401.87</td>\n",
       "      <td>14.62</td>\n",
       "      <td>...</td>\n",
       "      <td>2388.08</td>\n",
       "      <td>8133.83</td>\n",
       "      <td>8.3682</td>\n",
       "      <td>0.03</td>\n",
       "      <td>392</td>\n",
       "      <td>2388</td>\n",
       "      <td>100</td>\n",
       "      <td>38.88</td>\n",
       "      <td>23.3739</td>\n",
       "      <td>98</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>-0.0019</td>\n",
       "      <td>-0.0002</td>\n",
       "      <td>100</td>\n",
       "      <td>518.67</td>\n",
       "      <td>642.37</td>\n",
       "      <td>1582.85</td>\n",
       "      <td>1406.22</td>\n",
       "      <td>14.62</td>\n",
       "      <td>...</td>\n",
       "      <td>2388.04</td>\n",
       "      <td>8133.80</td>\n",
       "      <td>8.4294</td>\n",
       "      <td>0.03</td>\n",
       "      <td>393</td>\n",
       "      <td>2388</td>\n",
       "      <td>100</td>\n",
       "      <td>38.90</td>\n",
       "      <td>23.4044</td>\n",
       "      <td>98</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20426</th>\n",
       "      <td>99</td>\n",
       "      <td>181</td>\n",
       "      <td>-0.0015</td>\n",
       "      <td>-0.0001</td>\n",
       "      <td>100</td>\n",
       "      <td>518.67</td>\n",
       "      <td>643.72</td>\n",
       "      <td>1600.39</td>\n",
       "      <td>1428.03</td>\n",
       "      <td>14.62</td>\n",
       "      <td>...</td>\n",
       "      <td>2388.29</td>\n",
       "      <td>8123.55</td>\n",
       "      <td>8.4885</td>\n",
       "      <td>0.03</td>\n",
       "      <td>396</td>\n",
       "      <td>2388</td>\n",
       "      <td>100</td>\n",
       "      <td>38.50</td>\n",
       "      <td>23.0425</td>\n",
       "      <td>20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20427</th>\n",
       "      <td>99</td>\n",
       "      <td>182</td>\n",
       "      <td>-0.0027</td>\n",
       "      <td>-0.0003</td>\n",
       "      <td>100</td>\n",
       "      <td>518.67</td>\n",
       "      <td>643.52</td>\n",
       "      <td>1605.33</td>\n",
       "      <td>1430.32</td>\n",
       "      <td>14.62</td>\n",
       "      <td>...</td>\n",
       "      <td>2388.27</td>\n",
       "      <td>8130.99</td>\n",
       "      <td>8.5124</td>\n",
       "      <td>0.03</td>\n",
       "      <td>393</td>\n",
       "      <td>2388</td>\n",
       "      <td>100</td>\n",
       "      <td>38.39</td>\n",
       "      <td>22.9674</td>\n",
       "      <td>20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20428</th>\n",
       "      <td>99</td>\n",
       "      <td>183</td>\n",
       "      <td>-0.0031</td>\n",
       "      <td>-0.0003</td>\n",
       "      <td>100</td>\n",
       "      <td>518.67</td>\n",
       "      <td>643.96</td>\n",
       "      <td>1606.95</td>\n",
       "      <td>1427.90</td>\n",
       "      <td>14.62</td>\n",
       "      <td>...</td>\n",
       "      <td>2388.31</td>\n",
       "      <td>8126.90</td>\n",
       "      <td>8.5374</td>\n",
       "      <td>0.03</td>\n",
       "      <td>395</td>\n",
       "      <td>2388</td>\n",
       "      <td>100</td>\n",
       "      <td>38.57</td>\n",
       "      <td>23.1440</td>\n",
       "      <td>20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20429</th>\n",
       "      <td>99</td>\n",
       "      <td>184</td>\n",
       "      <td>-0.0010</td>\n",
       "      <td>-0.0001</td>\n",
       "      <td>100</td>\n",
       "      <td>518.67</td>\n",
       "      <td>644.10</td>\n",
       "      <td>1600.20</td>\n",
       "      <td>1436.54</td>\n",
       "      <td>14.62</td>\n",
       "      <td>...</td>\n",
       "      <td>2388.33</td>\n",
       "      <td>8125.66</td>\n",
       "      <td>8.5592</td>\n",
       "      <td>0.03</td>\n",
       "      <td>395</td>\n",
       "      <td>2388</td>\n",
       "      <td>100</td>\n",
       "      <td>38.45</td>\n",
       "      <td>23.0478</td>\n",
       "      <td>20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20430</th>\n",
       "      <td>99</td>\n",
       "      <td>185</td>\n",
       "      <td>-0.0019</td>\n",
       "      <td>-0.0004</td>\n",
       "      <td>100</td>\n",
       "      <td>518.67</td>\n",
       "      <td>643.93</td>\n",
       "      <td>1598.42</td>\n",
       "      <td>1421.56</td>\n",
       "      <td>14.62</td>\n",
       "      <td>...</td>\n",
       "      <td>2388.24</td>\n",
       "      <td>8127.53</td>\n",
       "      <td>8.5425</td>\n",
       "      <td>0.03</td>\n",
       "      <td>397</td>\n",
       "      <td>2388</td>\n",
       "      <td>100</td>\n",
       "      <td>38.49</td>\n",
       "      <td>23.1931</td>\n",
       "      <td>20</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>20431 rows × 27 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       id  cycle  setting1  setting2  setting3      s1      s2       s3  \\\n",
       "0       1      1   -0.0007   -0.0004       100  518.67  641.82  1589.70   \n",
       "1       1      2    0.0019   -0.0003       100  518.67  642.15  1591.82   \n",
       "2       1      3   -0.0043    0.0003       100  518.67  642.35  1587.99   \n",
       "3       1      4    0.0007    0.0000       100  518.67  642.35  1582.79   \n",
       "4       1      5   -0.0019   -0.0002       100  518.67  642.37  1582.85   \n",
       "...    ..    ...       ...       ...       ...     ...     ...      ...   \n",
       "20426  99    181   -0.0015   -0.0001       100  518.67  643.72  1600.39   \n",
       "20427  99    182   -0.0027   -0.0003       100  518.67  643.52  1605.33   \n",
       "20428  99    183   -0.0031   -0.0003       100  518.67  643.96  1606.95   \n",
       "20429  99    184   -0.0010   -0.0001       100  518.67  644.10  1600.20   \n",
       "20430  99    185   -0.0019   -0.0004       100  518.67  643.93  1598.42   \n",
       "\n",
       "            s4     s5  ...      s13      s14     s15   s16  s17   s18  s19  \\\n",
       "0      1400.60  14.62  ...  2388.02  8138.62  8.4195  0.03  392  2388  100   \n",
       "1      1403.14  14.62  ...  2388.07  8131.49  8.4318  0.03  392  2388  100   \n",
       "2      1404.20  14.62  ...  2388.03  8133.23  8.4178  0.03  390  2388  100   \n",
       "3      1401.87  14.62  ...  2388.08  8133.83  8.3682  0.03  392  2388  100   \n",
       "4      1406.22  14.62  ...  2388.04  8133.80  8.4294  0.03  393  2388  100   \n",
       "...        ...    ...  ...      ...      ...     ...   ...  ...   ...  ...   \n",
       "20426  1428.03  14.62  ...  2388.29  8123.55  8.4885  0.03  396  2388  100   \n",
       "20427  1430.32  14.62  ...  2388.27  8130.99  8.5124  0.03  393  2388  100   \n",
       "20428  1427.90  14.62  ...  2388.31  8126.90  8.5374  0.03  395  2388  100   \n",
       "20429  1436.54  14.62  ...  2388.33  8125.66  8.5592  0.03  395  2388  100   \n",
       "20430  1421.56  14.62  ...  2388.24  8127.53  8.5425  0.03  397  2388  100   \n",
       "\n",
       "         s20      s21  more  \n",
       "0      39.06  23.4190    98  \n",
       "1      39.00  23.4236    98  \n",
       "2      38.95  23.3442    98  \n",
       "3      38.88  23.3739    98  \n",
       "4      38.90  23.4044    98  \n",
       "...      ...      ...   ...  \n",
       "20426  38.50  23.0425    20  \n",
       "20427  38.39  22.9674    20  \n",
       "20428  38.57  23.1440    20  \n",
       "20429  38.45  23.0478    20  \n",
       "20430  38.49  23.1931    20  \n",
       "\n",
       "[20431 rows x 27 columns]"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_merged = pd.merge(df, df_truth, on='id')\n",
    "df_merged"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Step 1: Get the maximum cycle for each engine\n",
    "max_cycle_per_engine = df.groupby('id')['cycle'].max().reset_index()\n",
    "max_cycle_per_engine.columns = ['id', 'max_cycle']\n",
    "\n",
    "# Step 2: Merge the maximum cycle with the df_truth to get the actual failure cycle\n",
    "df_merged = pd.merge(max_cycle_per_engine, df_truth, on='id')\n",
    "\n",
    "# Step 3: Calculate the actual failure cycle (when engine will fail)\n",
    "df_merged['failure_cycle'] = df_merged['max_cycle'] + df_merged['more']\n",
    "\n",
    "# Step 4: Merge this back with the main DataFrame to compute remaining cycles\n",
    "df = pd.merge(df, df_merged[['id', 'failure_cycle']], on='id')\n",
    "\n",
    "# Step 5: Calculate remaining cycles for each row by subtracting the current cycle from the failure cycle\n",
    "df['remaining_cycles'] = df['failure_cycle'] - df['cycle']\n",
    "df = df.drop('failure_cycle',axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>cycle</th>\n",
       "      <th>setting1</th>\n",
       "      <th>setting2</th>\n",
       "      <th>setting3</th>\n",
       "      <th>s1</th>\n",
       "      <th>s2</th>\n",
       "      <th>s3</th>\n",
       "      <th>s4</th>\n",
       "      <th>s5</th>\n",
       "      <th>...</th>\n",
       "      <th>s13</th>\n",
       "      <th>s14</th>\n",
       "      <th>s15</th>\n",
       "      <th>s16</th>\n",
       "      <th>s17</th>\n",
       "      <th>s18</th>\n",
       "      <th>s19</th>\n",
       "      <th>s20</th>\n",
       "      <th>s21</th>\n",
       "      <th>remaining_cycles</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>-0.0007</td>\n",
       "      <td>-0.0004</td>\n",
       "      <td>100</td>\n",
       "      <td>518.67</td>\n",
       "      <td>641.82</td>\n",
       "      <td>1589.70</td>\n",
       "      <td>1400.60</td>\n",
       "      <td>14.62</td>\n",
       "      <td>...</td>\n",
       "      <td>2388.02</td>\n",
       "      <td>8138.62</td>\n",
       "      <td>8.4195</td>\n",
       "      <td>0.03</td>\n",
       "      <td>392</td>\n",
       "      <td>2388</td>\n",
       "      <td>100</td>\n",
       "      <td>39.06</td>\n",
       "      <td>23.4190</td>\n",
       "      <td>289</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0.0019</td>\n",
       "      <td>-0.0003</td>\n",
       "      <td>100</td>\n",
       "      <td>518.67</td>\n",
       "      <td>642.15</td>\n",
       "      <td>1591.82</td>\n",
       "      <td>1403.14</td>\n",
       "      <td>14.62</td>\n",
       "      <td>...</td>\n",
       "      <td>2388.07</td>\n",
       "      <td>8131.49</td>\n",
       "      <td>8.4318</td>\n",
       "      <td>0.03</td>\n",
       "      <td>392</td>\n",
       "      <td>2388</td>\n",
       "      <td>100</td>\n",
       "      <td>39.00</td>\n",
       "      <td>23.4236</td>\n",
       "      <td>288</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>-0.0043</td>\n",
       "      <td>0.0003</td>\n",
       "      <td>100</td>\n",
       "      <td>518.67</td>\n",
       "      <td>642.35</td>\n",
       "      <td>1587.99</td>\n",
       "      <td>1404.20</td>\n",
       "      <td>14.62</td>\n",
       "      <td>...</td>\n",
       "      <td>2388.03</td>\n",
       "      <td>8133.23</td>\n",
       "      <td>8.4178</td>\n",
       "      <td>0.03</td>\n",
       "      <td>390</td>\n",
       "      <td>2388</td>\n",
       "      <td>100</td>\n",
       "      <td>38.95</td>\n",
       "      <td>23.3442</td>\n",
       "      <td>287</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "      <td>0.0007</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>100</td>\n",
       "      <td>518.67</td>\n",
       "      <td>642.35</td>\n",
       "      <td>1582.79</td>\n",
       "      <td>1401.87</td>\n",
       "      <td>14.62</td>\n",
       "      <td>...</td>\n",
       "      <td>2388.08</td>\n",
       "      <td>8133.83</td>\n",
       "      <td>8.3682</td>\n",
       "      <td>0.03</td>\n",
       "      <td>392</td>\n",
       "      <td>2388</td>\n",
       "      <td>100</td>\n",
       "      <td>38.88</td>\n",
       "      <td>23.3739</td>\n",
       "      <td>286</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>-0.0019</td>\n",
       "      <td>-0.0002</td>\n",
       "      <td>100</td>\n",
       "      <td>518.67</td>\n",
       "      <td>642.37</td>\n",
       "      <td>1582.85</td>\n",
       "      <td>1406.22</td>\n",
       "      <td>14.62</td>\n",
       "      <td>...</td>\n",
       "      <td>2388.04</td>\n",
       "      <td>8133.80</td>\n",
       "      <td>8.4294</td>\n",
       "      <td>0.03</td>\n",
       "      <td>393</td>\n",
       "      <td>2388</td>\n",
       "      <td>100</td>\n",
       "      <td>38.90</td>\n",
       "      <td>23.4044</td>\n",
       "      <td>285</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 27 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   id  cycle  setting1  setting2  setting3      s1      s2       s3       s4  \\\n",
       "0   1      1   -0.0007   -0.0004       100  518.67  641.82  1589.70  1400.60   \n",
       "1   1      2    0.0019   -0.0003       100  518.67  642.15  1591.82  1403.14   \n",
       "2   1      3   -0.0043    0.0003       100  518.67  642.35  1587.99  1404.20   \n",
       "3   1      4    0.0007    0.0000       100  518.67  642.35  1582.79  1401.87   \n",
       "4   1      5   -0.0019   -0.0002       100  518.67  642.37  1582.85  1406.22   \n",
       "\n",
       "      s5  ...      s13      s14     s15   s16  s17   s18  s19    s20      s21  \\\n",
       "0  14.62  ...  2388.02  8138.62  8.4195  0.03  392  2388  100  39.06  23.4190   \n",
       "1  14.62  ...  2388.07  8131.49  8.4318  0.03  392  2388  100  39.00  23.4236   \n",
       "2  14.62  ...  2388.03  8133.23  8.4178  0.03  390  2388  100  38.95  23.3442   \n",
       "3  14.62  ...  2388.08  8133.83  8.3682  0.03  392  2388  100  38.88  23.3739   \n",
       "4  14.62  ...  2388.04  8133.80  8.4294  0.03  393  2388  100  38.90  23.4044   \n",
       "\n",
       "   remaining_cycles  \n",
       "0               289  \n",
       "1               288  \n",
       "2               287  \n",
       "3               286  \n",
       "4               285  \n",
       "\n",
       "[5 rows x 27 columns]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 20431 entries, 0 to 20430\n",
      "Data columns (total 27 columns):\n",
      " #   Column            Non-Null Count  Dtype  \n",
      "---  ------            --------------  -----  \n",
      " 0   id                20431 non-null  int64  \n",
      " 1   cycle             20431 non-null  int64  \n",
      " 2   setting1          20431 non-null  float64\n",
      " 3   setting2          20431 non-null  float64\n",
      " 4   setting3          20431 non-null  int64  \n",
      " 5   s1                20431 non-null  float64\n",
      " 6   s2                20431 non-null  float64\n",
      " 7   s3                20431 non-null  float64\n",
      " 8   s4                20431 non-null  float64\n",
      " 9   s5                20431 non-null  float64\n",
      " 10  s6                20431 non-null  float64\n",
      " 11  s7                20431 non-null  float64\n",
      " 12  s8                20431 non-null  float64\n",
      " 13  s9                20431 non-null  float64\n",
      " 14  s10               20431 non-null  float64\n",
      " 15  s11               20431 non-null  float64\n",
      " 16  s12               20431 non-null  float64\n",
      " 17  s13               20431 non-null  float64\n",
      " 18  s14               20431 non-null  float64\n",
      " 19  s15               20431 non-null  float64\n",
      " 20  s16               20431 non-null  float64\n",
      " 21  s17               20431 non-null  int64  \n",
      " 22  s18               20431 non-null  int64  \n",
      " 23  s19               20431 non-null  int64  \n",
      " 24  s20               20431 non-null  float64\n",
      " 25  s21               20431 non-null  float64\n",
      " 26  remaining_cycles  20431 non-null  int64  \n",
      "dtypes: float64(20), int64(7)\n",
      "memory usage: 4.2 MB\n"
     ]
    }
   ],
   "source": [
    "df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "def create_sequences_3(df, window_size=30, test_size=0.1, val_size=0.1):\n",
    "    \"\"\"\n",
    "    Generates scaled sequences and splits into training, validation, and testing sets.\n",
    "    \n",
    "    Parameters:\n",
    "    df (pd.DataFrame): The input dataframe with engine data.\n",
    "    window_size (int): Number of time steps in each sequence.\n",
    "    test_size (float): Fraction of data to reserve for testing.\n",
    "    val_size (float): Fraction of data to reserve for validation.\n",
    "\n",
    "    Returns:\n",
    "    X_train, X_val, X_test, y_train, y_val, y_test, feature_scaler, target_scaler (np.ndarray): \n",
    "    Arrays of train/val/test sequences and targets, feature and target scalers.\n",
    "    \"\"\"\n",
    "    features = [col for col in df.columns if col not in ['id', 'cycle', 'remaining_cycles']]\n",
    "    target_column = 'remaining_cycles'\n",
    "    \n",
    "    # Initialize scalers\n",
    "    feature_scaler = StandardScaler()\n",
    "    target_scaler = StandardScaler()\n",
    "    \n",
    "    # Lists to collect sequences and targets\n",
    "    X_sequences = []\n",
    "    y_targets = []\n",
    "    \n",
    "    # Group by engine ID and create sliding windows before splitting\n",
    "    for engine_id, engine_data in df.groupby('id'):\n",
    "        engine_data = engine_data.sort_values(by='cycle')  # Sort by cycle to keep temporal order\n",
    "        \n",
    "        # Create sliding windows\n",
    "        for i in range(len(engine_data) - window_size):\n",
    "            X_sequence = engine_data[features].iloc[i:i + window_size].values\n",
    "            y_target = engine_data[target_column].iloc[i + window_size - 1]  # Target is RUL of last cycle in window\n",
    "            \n",
    "            X_sequences.append(X_sequence)\n",
    "            y_targets.append(y_target)\n",
    "    \n",
    "    # Convert to numpy arrays\n",
    "    X_sequences = np.array(X_sequences)\n",
    "    y_targets = np.array(y_targets)\n",
    "    \n",
    "    # Split data into training+validation and testing\n",
    "    X_train_val, X_test, y_train_val, y_test = train_test_split(X_sequences, y_targets, test_size=test_size, random_state=42)\n",
    "    \n",
    "    # Further split the training+validation data into training and validation\n",
    "    val_ratio_adjusted = val_size / (1 - test_size)  # Adjust validation size proportionally\n",
    "    X_train, X_val, y_train, y_val = train_test_split(X_train_val, y_train_val, test_size=val_ratio_adjusted, random_state=42)\n",
    "    \n",
    "    # Fit the scalers on the training data\n",
    "    feature_scaler.fit(X_train.reshape(-1, X_train.shape[2]))  # Reshape for fitting scaler\n",
    "    target_scaler.fit(y_train.reshape(-1, 1))  # Reshape target for fitting scaler\n",
    "    \n",
    "    # Apply the scalers to the data\n",
    "    X_train = feature_scaler.transform(X_train.reshape(-1, X_train.shape[2])).reshape(X_train.shape)  # Reshape back after scaling\n",
    "    X_val = feature_scaler.transform(X_val.reshape(-1, X_val.shape[2])).reshape(X_val.shape)\n",
    "    X_test = feature_scaler.transform(X_test.reshape(-1, X_test.shape[2])).reshape(X_test.shape)\n",
    "    \n",
    "    y_train = target_scaler.transform(y_train.reshape(-1, 1))\n",
    "    y_val = target_scaler.transform(y_val.reshape(-1, 1))\n",
    "    y_test = target_scaler.transform(y_test.reshape(-1, 1))\n",
    "    \n",
    "    return X_train, X_val, X_test, y_train, y_val, y_test, feature_scaler, target_scaler\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_loss(history, title,lim=(None,None)):\n",
    "    plt.plot(history.history['loss'], label='Training Loss')\n",
    "    plt.plot(history.history['val_loss'], label='Validation Loss')\n",
    "    plt.xlabel('Epochs')\n",
    "    plt.ylabel('Loss')\n",
    "    plt.title(title)\n",
    "    plt.legend()\n",
    "    plt.ylim(lim)\n",
    "    plt.grid(True)\n",
    "    plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Adjusted evaluation function\n",
    "def evaluate_model(model, X_test, y_test,type,scaler):\n",
    "    predictions = model.predict(X_test)\n",
    "    if predictions.shape[0] != y_test.shape[0]:\n",
    "        # Trimming to the minimum size for consistency\n",
    "        min_len = min(predictions.shape[0], y_test.shape[0])\n",
    "        predictions = predictions[:min_len]\n",
    "        y_test = y_test[:min_len]\n",
    "    rmse = np.sqrt(mean_squared_error(scaler.inverse_transform(y_test.reshape(-1,1)), scaler.inverse_transform(predictions.reshape(-1,1))))\n",
    "    mae = mean_absolute_error(scaler.inverse_transform(y_test.reshape(-1,1)), scaler.inverse_transform(predictions.reshape(-1,1)))\n",
    "    print(f\"{type} Model Evaluation - RMSE: {rmse:.2f}, MAE: {mae:.2f}\")\n",
    "    return predictions, rmse, mae"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "window_size_1=5\n",
    "window_size_2=10\n",
    "window_size_3=20\n",
    "window_size_4=30\n",
    "window_size_5=40\n",
    "window_size_6=50\n",
    "metrics_results = []\n",
    "num_features=24"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/200\n",
      "873/873 [==============================] - 20s 15ms/step - loss: 0.6123 - mse: 0.5857 - val_loss: 0.5942 - val_mse: 0.5600\n",
      "Epoch 2/200\n",
      "873/873 [==============================] - 12s 14ms/step - loss: 0.5988 - mse: 0.5672 - val_loss: 0.5815 - val_mse: 0.5313\n",
      "Epoch 3/200\n",
      "873/873 [==============================] - 11s 13ms/step - loss: 0.5967 - mse: 0.5620 - val_loss: 0.5797 - val_mse: 0.5313\n",
      "Epoch 4/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.5915 - mse: 0.5548 - val_loss: 0.5842 - val_mse: 0.5339\n",
      "Epoch 5/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.5875 - mse: 0.5498 - val_loss: 0.5807 - val_mse: 0.5179\n",
      "Epoch 6/200\n",
      "873/873 [==============================] - 11s 13ms/step - loss: 0.5856 - mse: 0.5469 - val_loss: 0.5675 - val_mse: 0.5139\n",
      "Epoch 7/200\n",
      "873/873 [==============================] - 11s 13ms/step - loss: 0.5781 - mse: 0.5397 - val_loss: 0.5680 - val_mse: 0.5198\n",
      "Epoch 8/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.5729 - mse: 0.5322 - val_loss: 0.5619 - val_mse: 0.5074\n",
      "Epoch 9/200\n",
      "873/873 [==============================] - 11s 13ms/step - loss: 0.5659 - mse: 0.5258 - val_loss: 0.5588 - val_mse: 0.5098\n",
      "Epoch 10/200\n",
      "873/873 [==============================] - 11s 13ms/step - loss: 0.5594 - mse: 0.5171 - val_loss: 0.5560 - val_mse: 0.5081\n",
      "Epoch 11/200\n",
      "873/873 [==============================] - 11s 13ms/step - loss: 0.5496 - mse: 0.5055 - val_loss: 0.5540 - val_mse: 0.5052\n",
      "Epoch 12/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.5425 - mse: 0.4967 - val_loss: 0.5458 - val_mse: 0.4904\n",
      "Epoch 13/200\n",
      "873/873 [==============================] - 11s 13ms/step - loss: 0.5338 - mse: 0.4856 - val_loss: 0.5417 - val_mse: 0.4862\n",
      "Epoch 14/200\n",
      "873/873 [==============================] - 11s 13ms/step - loss: 0.5251 - mse: 0.4745 - val_loss: 0.5416 - val_mse: 0.4941\n",
      "Epoch 15/200\n",
      "873/873 [==============================] - 11s 13ms/step - loss: 0.5186 - mse: 0.4636 - val_loss: 0.5307 - val_mse: 0.4684\n",
      "Epoch 16/200\n",
      "873/873 [==============================] - 11s 13ms/step - loss: 0.5092 - mse: 0.4521 - val_loss: 0.5417 - val_mse: 0.5060\n",
      "Epoch 17/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.4962 - mse: 0.4339 - val_loss: 0.5339 - val_mse: 0.4881\n",
      "Epoch 18/200\n",
      "873/873 [==============================] - 11s 13ms/step - loss: 0.4887 - mse: 0.4226 - val_loss: 0.5255 - val_mse: 0.4622\n",
      "Epoch 19/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.4780 - mse: 0.4094 - val_loss: 0.5182 - val_mse: 0.4606\n",
      "Epoch 20/200\n",
      "873/873 [==============================] - 11s 13ms/step - loss: 0.4692 - mse: 0.3961 - val_loss: 0.5278 - val_mse: 0.4769\n",
      "Epoch 21/200\n",
      "873/873 [==============================] - 11s 13ms/step - loss: 0.4607 - mse: 0.3832 - val_loss: 0.5110 - val_mse: 0.4487\n",
      "Epoch 22/200\n",
      "873/873 [==============================] - 11s 13ms/step - loss: 0.4563 - mse: 0.3780 - val_loss: 0.5099 - val_mse: 0.4448\n",
      "Epoch 23/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.4460 - mse: 0.3606 - val_loss: 0.5111 - val_mse: 0.4573\n",
      "Epoch 24/200\n",
      "873/873 [==============================] - 13s 14ms/step - loss: 0.4418 - mse: 0.3560 - val_loss: 0.5059 - val_mse: 0.4314\n",
      "Epoch 25/200\n",
      "873/873 [==============================] - 11s 13ms/step - loss: 0.4343 - mse: 0.3476 - val_loss: 0.5003 - val_mse: 0.4321\n",
      "Epoch 26/200\n",
      "873/873 [==============================] - 11s 13ms/step - loss: 0.4273 - mse: 0.3349 - val_loss: 0.5041 - val_mse: 0.4361\n",
      "Epoch 27/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.4203 - mse: 0.3241 - val_loss: 0.4973 - val_mse: 0.4254\n",
      "Epoch 28/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.4154 - mse: 0.3160 - val_loss: 0.4950 - val_mse: 0.4237\n",
      "Epoch 29/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.4088 - mse: 0.3095 - val_loss: 0.4912 - val_mse: 0.4154\n",
      "Epoch 30/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.4059 - mse: 0.3035 - val_loss: 0.4886 - val_mse: 0.4076\n",
      "Epoch 31/200\n",
      "873/873 [==============================] - 10s 12ms/step - loss: 0.4008 - mse: 0.2994 - val_loss: 0.4838 - val_mse: 0.4082\n",
      "Epoch 32/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.3954 - mse: 0.2902 - val_loss: 0.4916 - val_mse: 0.4168\n",
      "Epoch 33/200\n",
      "873/873 [==============================] - 10s 12ms/step - loss: 0.3911 - mse: 0.2841 - val_loss: 0.4864 - val_mse: 0.4088\n",
      "Epoch 34/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.3826 - mse: 0.2733 - val_loss: 0.4822 - val_mse: 0.4048\n",
      "Epoch 35/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.3793 - mse: 0.2673 - val_loss: 0.4811 - val_mse: 0.4009\n",
      "Epoch 36/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.3748 - mse: 0.2609 - val_loss: 0.4881 - val_mse: 0.4114\n",
      "Epoch 37/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.3714 - mse: 0.2568 - val_loss: 0.4754 - val_mse: 0.3911\n",
      "Epoch 38/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.3692 - mse: 0.2539 - val_loss: 0.4736 - val_mse: 0.3922\n",
      "Epoch 39/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.3629 - mse: 0.2437 - val_loss: 0.4766 - val_mse: 0.3900\n",
      "Epoch 40/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.3623 - mse: 0.2436 - val_loss: 0.4875 - val_mse: 0.4105\n",
      "Epoch 41/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.3577 - mse: 0.2387 - val_loss: 0.4747 - val_mse: 0.3910\n",
      "Epoch 42/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.3553 - mse: 0.2346 - val_loss: 0.4728 - val_mse: 0.3915\n",
      "Epoch 43/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.3524 - mse: 0.2329 - val_loss: 0.4884 - val_mse: 0.4084\n",
      "Epoch 44/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.3494 - mse: 0.2296 - val_loss: 0.4734 - val_mse: 0.3887\n",
      "Epoch 45/200\n",
      "873/873 [==============================] - 11s 13ms/step - loss: 0.3436 - mse: 0.2223 - val_loss: 0.5031 - val_mse: 0.4325\n",
      "Epoch 46/200\n",
      "873/873 [==============================] - 11s 13ms/step - loss: 0.3441 - mse: 0.2204 - val_loss: 0.4775 - val_mse: 0.3911\n",
      "Epoch 47/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.3432 - mse: 0.2186 - val_loss: 0.4799 - val_mse: 0.3943\n",
      "Epoch 48/200\n",
      "873/873 [==============================] - 10s 12ms/step - loss: 0.3397 - mse: 0.2139 - val_loss: 0.4753 - val_mse: 0.3881\n",
      "Epoch 49/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.3337 - mse: 0.2081 - val_loss: 0.4689 - val_mse: 0.3840\n",
      "Epoch 50/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.3344 - mse: 0.2073 - val_loss: 0.4705 - val_mse: 0.3892\n",
      "Epoch 51/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.3309 - mse: 0.2042 - val_loss: 0.4757 - val_mse: 0.3861\n",
      "Epoch 52/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.3293 - mse: 0.2020 - val_loss: 0.4685 - val_mse: 0.3798\n",
      "Epoch 53/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.3247 - mse: 0.1961 - val_loss: 0.4624 - val_mse: 0.3744\n",
      "Epoch 54/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.3237 - mse: 0.1952 - val_loss: 0.4711 - val_mse: 0.3855\n",
      "Epoch 55/200\n",
      "873/873 [==============================] - 10s 12ms/step - loss: 0.3226 - mse: 0.1936 - val_loss: 0.4726 - val_mse: 0.3869\n",
      "Epoch 56/200\n",
      "873/873 [==============================] - 10s 12ms/step - loss: 0.3195 - mse: 0.1906 - val_loss: 0.4761 - val_mse: 0.3939\n",
      "Epoch 57/200\n",
      "873/873 [==============================] - 10s 12ms/step - loss: 0.3194 - mse: 0.1885 - val_loss: 0.4666 - val_mse: 0.3768\n",
      "Epoch 58/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.3132 - mse: 0.1846 - val_loss: 0.4734 - val_mse: 0.3902\n",
      "Epoch 59/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.3147 - mse: 0.1848 - val_loss: 0.4682 - val_mse: 0.3809\n",
      "Epoch 60/200\n",
      "873/873 [==============================] - 10s 12ms/step - loss: 0.3141 - mse: 0.1832 - val_loss: 0.4722 - val_mse: 0.3841\n",
      "Epoch 61/200\n",
      "873/873 [==============================] - 10s 12ms/step - loss: 0.3099 - mse: 0.1789 - val_loss: 0.4718 - val_mse: 0.3842\n",
      "Epoch 62/200\n",
      "873/873 [==============================] - 10s 12ms/step - loss: 0.3089 - mse: 0.1780 - val_loss: 0.4684 - val_mse: 0.3803\n",
      "Epoch 63/200\n",
      "873/873 [==============================] - 10s 12ms/step - loss: 0.3109 - mse: 0.1782 - val_loss: 0.4672 - val_mse: 0.3821\n",
      "Epoch 64/200\n",
      "873/873 [==============================] - 10s 12ms/step - loss: 0.3076 - mse: 0.1737 - val_loss: 0.4701 - val_mse: 0.3811\n",
      "Epoch 65/200\n",
      "873/873 [==============================] - 10s 12ms/step - loss: 0.3033 - mse: 0.1695 - val_loss: 0.4681 - val_mse: 0.3808\n",
      "Epoch 66/200\n",
      "873/873 [==============================] - 10s 11ms/step - loss: 0.3006 - mse: 0.1680 - val_loss: 0.4687 - val_mse: 0.3793\n",
      "Epoch 67/200\n",
      "873/873 [==============================] - 10s 12ms/step - loss: 0.3005 - mse: 0.1678 - val_loss: 0.4678 - val_mse: 0.3866\n",
      "Epoch 68/200\n",
      "873/873 [==============================] - 10s 12ms/step - loss: 0.2995 - mse: 0.1669 - val_loss: 0.4721 - val_mse: 0.3883\n",
      "Epoch 69/200\n",
      "873/873 [==============================] - 10s 12ms/step - loss: 0.2992 - mse: 0.1671 - val_loss: 0.4718 - val_mse: 0.3870\n",
      "Epoch 70/200\n",
      "873/873 [==============================] - 10s 12ms/step - loss: 0.2955 - mse: 0.1624 - val_loss: 0.4751 - val_mse: 0.3910\n",
      "Epoch 71/200\n",
      "873/873 [==============================] - 10s 12ms/step - loss: 0.2938 - mse: 0.1598 - val_loss: 0.4597 - val_mse: 0.3742\n",
      "Epoch 72/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.2935 - mse: 0.1591 - val_loss: 0.4642 - val_mse: 0.3756\n",
      "Epoch 73/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.2944 - mse: 0.1586 - val_loss: 0.4668 - val_mse: 0.3765\n",
      "Epoch 74/200\n",
      "873/873 [==============================] - 10s 12ms/step - loss: 0.2905 - mse: 0.1569 - val_loss: 0.4665 - val_mse: 0.3807\n",
      "Epoch 75/200\n",
      "873/873 [==============================] - 10s 12ms/step - loss: 0.2890 - mse: 0.1556 - val_loss: 0.4685 - val_mse: 0.3864\n",
      "Epoch 76/200\n",
      "873/873 [==============================] - 10s 12ms/step - loss: 0.2883 - mse: 0.1533 - val_loss: 0.4689 - val_mse: 0.3825\n",
      "Epoch 77/200\n",
      "873/873 [==============================] - 10s 12ms/step - loss: 0.2863 - mse: 0.1523 - val_loss: 0.4621 - val_mse: 0.3726\n",
      "Epoch 78/200\n",
      "873/873 [==============================] - 10s 12ms/step - loss: 0.2842 - mse: 0.1483 - val_loss: 0.4768 - val_mse: 0.3922\n",
      "Epoch 79/200\n",
      "873/873 [==============================] - 10s 12ms/step - loss: 0.2849 - mse: 0.1485 - val_loss: 0.4721 - val_mse: 0.3874\n",
      "Epoch 80/200\n",
      "873/873 [==============================] - 10s 12ms/step - loss: 0.2828 - mse: 0.1476 - val_loss: 0.4695 - val_mse: 0.3844\n",
      "Epoch 81/200\n",
      "873/873 [==============================] - 10s 12ms/step - loss: 0.2811 - mse: 0.1469 - val_loss: 0.4723 - val_mse: 0.3855\n",
      "Epoch 82/200\n",
      "873/873 [==============================] - 10s 12ms/step - loss: 0.2795 - mse: 0.1439 - val_loss: 0.4653 - val_mse: 0.3788\n",
      "Epoch 83/200\n",
      "873/873 [==============================] - 10s 12ms/step - loss: 0.2818 - mse: 0.1453 - val_loss: 0.4672 - val_mse: 0.3822\n",
      "Epoch 84/200\n",
      "873/873 [==============================] - 10s 12ms/step - loss: 0.2810 - mse: 0.1456 - val_loss: 0.4642 - val_mse: 0.3752\n",
      "Epoch 85/200\n",
      "873/873 [==============================] - 10s 12ms/step - loss: 0.2767 - mse: 0.1417 - val_loss: 0.4717 - val_mse: 0.3892\n",
      "Epoch 86/200\n",
      "873/873 [==============================] - 10s 12ms/step - loss: 0.2770 - mse: 0.1408 - val_loss: 0.4677 - val_mse: 0.3788\n",
      "Epoch 87/200\n",
      "873/873 [==============================] - 10s 12ms/step - loss: 0.2763 - mse: 0.1416 - val_loss: 0.4645 - val_mse: 0.3752\n",
      "Epoch 88/200\n",
      "873/873 [==============================] - 10s 12ms/step - loss: 0.2736 - mse: 0.1394 - val_loss: 0.4608 - val_mse: 0.3709\n",
      "Epoch 89/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.2720 - mse: 0.1365 - val_loss: 0.4620 - val_mse: 0.3750\n",
      "Epoch 90/200\n",
      "873/873 [==============================] - 11s 13ms/step - loss: 0.2720 - mse: 0.1369 - val_loss: 0.4687 - val_mse: 0.3896\n",
      "Epoch 91/200\n",
      "873/873 [==============================] - 10s 12ms/step - loss: 0.2726 - mse: 0.1379 - val_loss: 0.4689 - val_mse: 0.3872\n",
      "Epoch 92/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.2715 - mse: 0.1368 - val_loss: 0.4616 - val_mse: 0.3767\n",
      "Epoch 93/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.2673 - mse: 0.1320 - val_loss: 0.4630 - val_mse: 0.3769\n",
      "Epoch 94/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.2693 - mse: 0.1336 - val_loss: 0.4656 - val_mse: 0.3830\n",
      "Epoch 95/200\n",
      "873/873 [==============================] - 11s 12ms/step - loss: 0.2694 - mse: 0.1333 - val_loss: 0.4694 - val_mse: 0.3883\n",
      "Epoch 96/200\n",
      "873/873 [==============================] - 11s 13ms/step - loss: 0.2670 - mse: 0.1306 - val_loss: 0.4698 - val_mse: 0.3825\n",
      "Epoch 97/200\n",
      "873/873 [==============================] - 13s 14ms/step - loss: 0.2686 - mse: 0.1329 - val_loss: 0.4641 - val_mse: 0.3793\n",
      "Epoch 98/200\n",
      "873/873 [==============================] - 13s 14ms/step - loss: 0.2704 - mse: 0.1338 - val_loss: 0.4737 - val_mse: 0.3896\n",
      "Epoch 99/200\n",
      "873/873 [==============================] - 13s 14ms/step - loss: 0.2650 - mse: 0.1297 - val_loss: 0.4692 - val_mse: 0.3854\n",
      "Epoch 100/200\n",
      "873/873 [==============================] - 12s 14ms/step - loss: 0.2644 - mse: 0.1283 - val_loss: 0.4781 - val_mse: 0.3903\n",
      "Epoch 101/200\n",
      "873/873 [==============================] - 12s 14ms/step - loss: 0.2651 - mse: 0.1290 - val_loss: 0.4719 - val_mse: 0.3849\n"
     ]
    }
   ],
   "source": [
    "X_train, X_val, X_test, y_train, y_val, y_test, scaler,yscaler= create_sequences_3(df, window_size=window_size_1, test_size=0.2)\n",
    "\n",
    "model_1 = Sequential([\n",
    "    GRU(64, return_sequences=True, input_shape=(window_size_1, num_features)),\n",
    "    Dropout(0.3),\n",
    "    GRU(32, return_sequences=True),\n",
    "    Dropout(0.2),\n",
    "    GRU(16, return_sequences=False),\n",
    "    Dropout(0.1),\n",
    "    Dense(32, activation='relu'),\n",
    "    Dense(1)  # Output layer for regression\n",
    "])\n",
    "model_1.compile(optimizer=Adam(learning_rate=0.001),loss='mae',metrics=['mse'])\n",
    "# Train the model\n",
    "history_1 = model_1.fit(X_train,y_train,\n",
    "                    validation_data=(X_val,y_val),\n",
    "                    epochs=200, batch_size=16,\n",
    "                    callbacks=tf.keras.callbacks.EarlyStopping(patience=30,monitor='val_loss',restore_best_weights=True))\n",
    "model_1.save('model_1.keras')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/200\n",
      "851/851 [==============================] - 20s 17ms/step - loss: 0.6121 - mse: 0.5846 - val_loss: 0.5953 - val_mse: 0.5612\n",
      "Epoch 2/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.5985 - mse: 0.5646 - val_loss: 0.5798 - val_mse: 0.5440\n",
      "Epoch 3/200\n",
      "851/851 [==============================] - 14s 16ms/step - loss: 0.5888 - mse: 0.5543 - val_loss: 0.5753 - val_mse: 0.5378\n",
      "Epoch 4/200\n",
      "851/851 [==============================] - 12s 15ms/step - loss: 0.5760 - mse: 0.5375 - val_loss: 0.5792 - val_mse: 0.5462\n",
      "Epoch 5/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.5596 - mse: 0.5141 - val_loss: 0.5517 - val_mse: 0.5018\n",
      "Epoch 6/200\n",
      "851/851 [==============================] - 12s 14ms/step - loss: 0.5414 - mse: 0.4880 - val_loss: 0.5434 - val_mse: 0.4949\n",
      "Epoch 7/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.5162 - mse: 0.4514 - val_loss: 0.5154 - val_mse: 0.4601\n",
      "Epoch 8/200\n",
      "851/851 [==============================] - 14s 17ms/step - loss: 0.4885 - mse: 0.4155 - val_loss: 0.4808 - val_mse: 0.4005\n",
      "Epoch 9/200\n",
      "851/851 [==============================] - 14s 16ms/step - loss: 0.4678 - mse: 0.3872 - val_loss: 0.4610 - val_mse: 0.3797\n",
      "Epoch 10/200\n",
      "851/851 [==============================] - 13s 16ms/step - loss: 0.4404 - mse: 0.3473 - val_loss: 0.4398 - val_mse: 0.3537\n",
      "Epoch 11/200\n",
      "851/851 [==============================] - 12s 14ms/step - loss: 0.4154 - mse: 0.3108 - val_loss: 0.4276 - val_mse: 0.3332\n",
      "Epoch 12/200\n",
      "851/851 [==============================] - 12s 14ms/step - loss: 0.3953 - mse: 0.2853 - val_loss: 0.4356 - val_mse: 0.3345\n",
      "Epoch 13/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.3785 - mse: 0.2634 - val_loss: 0.3908 - val_mse: 0.2844\n",
      "Epoch 14/200\n",
      "851/851 [==============================] - 13s 16ms/step - loss: 0.3605 - mse: 0.2362 - val_loss: 0.3759 - val_mse: 0.2619\n",
      "Epoch 15/200\n",
      "851/851 [==============================] - 13s 16ms/step - loss: 0.3419 - mse: 0.2152 - val_loss: 0.3676 - val_mse: 0.2575\n",
      "Epoch 16/200\n",
      "851/851 [==============================] - 12s 14ms/step - loss: 0.3286 - mse: 0.1991 - val_loss: 0.3474 - val_mse: 0.2245\n",
      "Epoch 17/200\n",
      "851/851 [==============================] - 12s 14ms/step - loss: 0.3135 - mse: 0.1800 - val_loss: 0.3583 - val_mse: 0.2272\n",
      "Epoch 18/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.3019 - mse: 0.1697 - val_loss: 0.3292 - val_mse: 0.2050\n",
      "Epoch 19/200\n",
      "851/851 [==============================] - 13s 16ms/step - loss: 0.2927 - mse: 0.1575 - val_loss: 0.3162 - val_mse: 0.1899\n",
      "Epoch 20/200\n",
      "851/851 [==============================] - 13s 16ms/step - loss: 0.2831 - mse: 0.1468 - val_loss: 0.3190 - val_mse: 0.1902\n",
      "Epoch 21/200\n",
      "851/851 [==============================] - 13s 16ms/step - loss: 0.2756 - mse: 0.1392 - val_loss: 0.3312 - val_mse: 0.1983\n",
      "Epoch 22/200\n",
      "851/851 [==============================] - 14s 17ms/step - loss: 0.2671 - mse: 0.1314 - val_loss: 0.3015 - val_mse: 0.1661\n",
      "Epoch 23/200\n",
      "851/851 [==============================] - 12s 14ms/step - loss: 0.2568 - mse: 0.1202 - val_loss: 0.2976 - val_mse: 0.1643\n",
      "Epoch 24/200\n",
      "851/851 [==============================] - 12s 14ms/step - loss: 0.2542 - mse: 0.1189 - val_loss: 0.2952 - val_mse: 0.1689\n",
      "Epoch 25/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.2460 - mse: 0.1118 - val_loss: 0.2821 - val_mse: 0.1514\n",
      "Epoch 26/200\n",
      "851/851 [==============================] - 13s 16ms/step - loss: 0.2421 - mse: 0.1080 - val_loss: 0.2793 - val_mse: 0.1460\n",
      "Epoch 27/200\n",
      "851/851 [==============================] - 15s 17ms/step - loss: 0.2362 - mse: 0.1019 - val_loss: 0.2897 - val_mse: 0.1529\n",
      "Epoch 28/200\n",
      "851/851 [==============================] - 14s 16ms/step - loss: 0.2308 - mse: 0.0968 - val_loss: 0.2697 - val_mse: 0.1406\n",
      "Epoch 29/200\n",
      "851/851 [==============================] - 13s 16ms/step - loss: 0.2278 - mse: 0.0948 - val_loss: 0.2666 - val_mse: 0.1381\n",
      "Epoch 30/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.2237 - mse: 0.0908 - val_loss: 0.2721 - val_mse: 0.1394\n",
      "Epoch 31/200\n",
      "851/851 [==============================] - 13s 16ms/step - loss: 0.2193 - mse: 0.0868 - val_loss: 0.2637 - val_mse: 0.1323\n",
      "Epoch 32/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.2155 - mse: 0.0838 - val_loss: 0.2624 - val_mse: 0.1299\n",
      "Epoch 33/200\n",
      "851/851 [==============================] - 12s 15ms/step - loss: 0.2125 - mse: 0.0821 - val_loss: 0.2684 - val_mse: 0.1352\n",
      "Epoch 34/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.2093 - mse: 0.0789 - val_loss: 0.2453 - val_mse: 0.1119\n",
      "Epoch 35/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.2066 - mse: 0.0757 - val_loss: 0.2474 - val_mse: 0.1166\n",
      "Epoch 36/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.2022 - mse: 0.0735 - val_loss: 0.2598 - val_mse: 0.1250\n",
      "Epoch 37/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.2001 - mse: 0.0721 - val_loss: 0.2405 - val_mse: 0.1098\n",
      "Epoch 38/200\n",
      "851/851 [==============================] - 12s 14ms/step - loss: 0.1957 - mse: 0.0685 - val_loss: 0.2420 - val_mse: 0.1136\n",
      "Epoch 39/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1971 - mse: 0.0685 - val_loss: 0.2440 - val_mse: 0.1124\n",
      "Epoch 40/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1931 - mse: 0.0665 - val_loss: 0.2412 - val_mse: 0.1117\n",
      "Epoch 41/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1892 - mse: 0.0641 - val_loss: 0.2402 - val_mse: 0.1107\n",
      "Epoch 42/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1877 - mse: 0.0629 - val_loss: 0.2315 - val_mse: 0.1025\n",
      "Epoch 43/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1875 - mse: 0.0625 - val_loss: 0.2419 - val_mse: 0.1111\n",
      "Epoch 44/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1825 - mse: 0.0602 - val_loss: 0.2388 - val_mse: 0.1091\n",
      "Epoch 45/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1815 - mse: 0.0582 - val_loss: 0.2338 - val_mse: 0.1042\n",
      "Epoch 46/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1803 - mse: 0.0573 - val_loss: 0.2365 - val_mse: 0.1081\n",
      "Epoch 47/200\n",
      "851/851 [==============================] - 13s 16ms/step - loss: 0.1783 - mse: 0.0566 - val_loss: 0.2266 - val_mse: 0.1007\n",
      "Epoch 48/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1769 - mse: 0.0559 - val_loss: 0.2296 - val_mse: 0.1004\n",
      "Epoch 49/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1756 - mse: 0.0541 - val_loss: 0.2397 - val_mse: 0.1100\n",
      "Epoch 50/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1757 - mse: 0.0550 - val_loss: 0.2384 - val_mse: 0.1069\n",
      "Epoch 51/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1731 - mse: 0.0529 - val_loss: 0.2253 - val_mse: 0.0998\n",
      "Epoch 52/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1721 - mse: 0.0531 - val_loss: 0.2264 - val_mse: 0.1022\n",
      "Epoch 53/200\n",
      "851/851 [==============================] - 14s 16ms/step - loss: 0.1696 - mse: 0.0513 - val_loss: 0.2211 - val_mse: 0.0968\n",
      "Epoch 54/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1668 - mse: 0.0489 - val_loss: 0.2318 - val_mse: 0.1067\n",
      "Epoch 55/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1667 - mse: 0.0496 - val_loss: 0.2349 - val_mse: 0.1057\n",
      "Epoch 56/200\n",
      "851/851 [==============================] - 12s 15ms/step - loss: 0.1638 - mse: 0.0474 - val_loss: 0.2195 - val_mse: 0.0966\n",
      "Epoch 57/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1630 - mse: 0.0468 - val_loss: 0.2241 - val_mse: 0.0949\n",
      "Epoch 58/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1634 - mse: 0.0477 - val_loss: 0.2169 - val_mse: 0.0917\n",
      "Epoch 59/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1615 - mse: 0.0462 - val_loss: 0.2159 - val_mse: 0.0933\n",
      "Epoch 60/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1620 - mse: 0.0466 - val_loss: 0.2381 - val_mse: 0.1130\n",
      "Epoch 61/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1601 - mse: 0.0455 - val_loss: 0.2136 - val_mse: 0.0904\n",
      "Epoch 62/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1594 - mse: 0.0448 - val_loss: 0.2161 - val_mse: 0.0933\n",
      "Epoch 63/200\n",
      "851/851 [==============================] - 12s 14ms/step - loss: 0.1580 - mse: 0.0443 - val_loss: 0.2188 - val_mse: 0.0971\n",
      "Epoch 64/200\n",
      "851/851 [==============================] - 12s 14ms/step - loss: 0.1577 - mse: 0.0440 - val_loss: 0.2184 - val_mse: 0.0965\n",
      "Epoch 65/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1548 - mse: 0.0422 - val_loss: 0.2145 - val_mse: 0.0947\n",
      "Epoch 66/200\n",
      "851/851 [==============================] - 15s 17ms/step - loss: 0.1539 - mse: 0.0417 - val_loss: 0.2403 - val_mse: 0.1175\n",
      "Epoch 67/200\n",
      "851/851 [==============================] - 14s 16ms/step - loss: 0.1535 - mse: 0.0413 - val_loss: 0.2120 - val_mse: 0.0921\n",
      "Epoch 68/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1534 - mse: 0.0420 - val_loss: 0.2159 - val_mse: 0.0920\n",
      "Epoch 69/200\n",
      "851/851 [==============================] - 12s 15ms/step - loss: 0.1542 - mse: 0.0418 - val_loss: 0.2127 - val_mse: 0.0909\n",
      "Epoch 70/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1527 - mse: 0.0413 - val_loss: 0.2166 - val_mse: 0.0917\n",
      "Epoch 71/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1511 - mse: 0.0404 - val_loss: 0.2142 - val_mse: 0.0928\n",
      "Epoch 72/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1511 - mse: 0.0407 - val_loss: 0.2121 - val_mse: 0.0920\n",
      "Epoch 73/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1498 - mse: 0.0397 - val_loss: 0.2144 - val_mse: 0.0924\n",
      "Epoch 74/200\n",
      "851/851 [==============================] - 12s 15ms/step - loss: 0.1476 - mse: 0.0387 - val_loss: 0.2117 - val_mse: 0.0934\n",
      "Epoch 75/200\n",
      "851/851 [==============================] - 12s 14ms/step - loss: 0.1465 - mse: 0.0380 - val_loss: 0.2099 - val_mse: 0.0888\n",
      "Epoch 76/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1483 - mse: 0.0391 - val_loss: 0.2098 - val_mse: 0.0907\n",
      "Epoch 77/200\n",
      "851/851 [==============================] - 12s 14ms/step - loss: 0.1444 - mse: 0.0372 - val_loss: 0.2146 - val_mse: 0.0933\n",
      "Epoch 78/200\n",
      "851/851 [==============================] - 12s 15ms/step - loss: 0.1453 - mse: 0.0379 - val_loss: 0.2145 - val_mse: 0.0932\n",
      "Epoch 79/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1466 - mse: 0.0381 - val_loss: 0.2078 - val_mse: 0.0896\n",
      "Epoch 80/200\n",
      "851/851 [==============================] - 11s 13ms/step - loss: 0.1448 - mse: 0.0370 - val_loss: 0.2142 - val_mse: 0.0921\n",
      "Epoch 81/200\n",
      "851/851 [==============================] - 12s 14ms/step - loss: 0.1441 - mse: 0.0368 - val_loss: 0.2266 - val_mse: 0.1017\n",
      "Epoch 82/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1440 - mse: 0.0365 - val_loss: 0.2183 - val_mse: 0.0951\n",
      "Epoch 83/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1417 - mse: 0.0354 - val_loss: 0.2051 - val_mse: 0.0866\n",
      "Epoch 84/200\n",
      "851/851 [==============================] - 12s 14ms/step - loss: 0.1428 - mse: 0.0363 - val_loss: 0.2079 - val_mse: 0.0899\n",
      "Epoch 85/200\n",
      "851/851 [==============================] - 12s 15ms/step - loss: 0.1419 - mse: 0.0356 - val_loss: 0.2109 - val_mse: 0.0899\n",
      "Epoch 86/200\n",
      "851/851 [==============================] - 12s 15ms/step - loss: 0.1410 - mse: 0.0354 - val_loss: 0.2118 - val_mse: 0.0932\n",
      "Epoch 87/200\n",
      "851/851 [==============================] - 12s 14ms/step - loss: 0.1426 - mse: 0.0359 - val_loss: 0.2056 - val_mse: 0.0865\n",
      "Epoch 88/200\n",
      "851/851 [==============================] - 12s 14ms/step - loss: 0.1421 - mse: 0.0356 - val_loss: 0.2187 - val_mse: 0.0935\n",
      "Epoch 89/200\n",
      "851/851 [==============================] - 12s 14ms/step - loss: 0.1406 - mse: 0.0349 - val_loss: 0.2140 - val_mse: 0.0912\n",
      "Epoch 90/200\n",
      "851/851 [==============================] - 12s 15ms/step - loss: 0.1400 - mse: 0.0350 - val_loss: 0.2115 - val_mse: 0.0943\n",
      "Epoch 91/200\n",
      "851/851 [==============================] - 11s 13ms/step - loss: 0.1403 - mse: 0.0349 - val_loss: 0.2154 - val_mse: 0.0940\n",
      "Epoch 92/200\n",
      "851/851 [==============================] - 12s 14ms/step - loss: 0.1364 - mse: 0.0329 - val_loss: 0.2130 - val_mse: 0.0929\n",
      "Epoch 93/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1392 - mse: 0.0344 - val_loss: 0.2071 - val_mse: 0.0898\n",
      "Epoch 94/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1366 - mse: 0.0333 - val_loss: 0.2252 - val_mse: 0.1039\n",
      "Epoch 95/200\n",
      "851/851 [==============================] - 12s 14ms/step - loss: 0.1376 - mse: 0.0335 - val_loss: 0.2180 - val_mse: 0.1004\n",
      "Epoch 96/200\n",
      "851/851 [==============================] - 12s 14ms/step - loss: 0.1372 - mse: 0.0332 - val_loss: 0.2076 - val_mse: 0.0895\n",
      "Epoch 97/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1359 - mse: 0.0325 - val_loss: 0.2070 - val_mse: 0.0917\n",
      "Epoch 98/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1361 - mse: 0.0331 - val_loss: 0.2129 - val_mse: 0.0932\n",
      "Epoch 99/200\n",
      "851/851 [==============================] - 12s 15ms/step - loss: 0.1365 - mse: 0.0329 - val_loss: 0.2058 - val_mse: 0.0892\n",
      "Epoch 100/200\n",
      "851/851 [==============================] - 12s 14ms/step - loss: 0.1338 - mse: 0.0316 - val_loss: 0.2135 - val_mse: 0.0940\n",
      "Epoch 101/200\n",
      "851/851 [==============================] - 12s 14ms/step - loss: 0.1352 - mse: 0.0323 - val_loss: 0.2092 - val_mse: 0.0920\n",
      "Epoch 102/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1340 - mse: 0.0315 - val_loss: 0.2121 - val_mse: 0.0906\n",
      "Epoch 103/200\n",
      "851/851 [==============================] - 12s 14ms/step - loss: 0.1342 - mse: 0.0321 - val_loss: 0.2124 - val_mse: 0.0929\n",
      "Epoch 104/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1341 - mse: 0.0316 - val_loss: 0.1982 - val_mse: 0.0825\n",
      "Epoch 105/200\n",
      "851/851 [==============================] - 12s 14ms/step - loss: 0.1333 - mse: 0.0314 - val_loss: 0.2041 - val_mse: 0.0877\n",
      "Epoch 106/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1335 - mse: 0.0315 - val_loss: 0.2076 - val_mse: 0.0911\n",
      "Epoch 107/200\n",
      "851/851 [==============================] - 13s 16ms/step - loss: 0.1300 - mse: 0.0298 - val_loss: 0.2107 - val_mse: 0.0946\n",
      "Epoch 108/200\n",
      "851/851 [==============================] - 13s 16ms/step - loss: 0.1298 - mse: 0.0301 - val_loss: 0.2178 - val_mse: 0.0976\n",
      "Epoch 109/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1293 - mse: 0.0294 - val_loss: 0.2070 - val_mse: 0.0889\n",
      "Epoch 110/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1326 - mse: 0.0310 - val_loss: 0.2051 - val_mse: 0.0906\n",
      "Epoch 111/200\n",
      "851/851 [==============================] - 13s 16ms/step - loss: 0.1322 - mse: 0.0305 - val_loss: 0.2135 - val_mse: 0.0970\n",
      "Epoch 112/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1306 - mse: 0.0300 - val_loss: 0.2168 - val_mse: 0.0962\n",
      "Epoch 113/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1302 - mse: 0.0301 - val_loss: 0.2085 - val_mse: 0.0927\n",
      "Epoch 114/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1295 - mse: 0.0296 - val_loss: 0.2062 - val_mse: 0.0883\n",
      "Epoch 115/200\n",
      "851/851 [==============================] - 13s 16ms/step - loss: 0.1296 - mse: 0.0298 - val_loss: 0.2123 - val_mse: 0.0939\n",
      "Epoch 116/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1291 - mse: 0.0293 - val_loss: 0.2029 - val_mse: 0.0894\n",
      "Epoch 117/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1282 - mse: 0.0292 - val_loss: 0.2091 - val_mse: 0.0930\n",
      "Epoch 118/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1288 - mse: 0.0295 - val_loss: 0.2102 - val_mse: 0.0948\n",
      "Epoch 119/200\n",
      "851/851 [==============================] - 14s 16ms/step - loss: 0.1276 - mse: 0.0293 - val_loss: 0.2102 - val_mse: 0.0932\n",
      "Epoch 120/200\n",
      "851/851 [==============================] - 12s 14ms/step - loss: 0.1284 - mse: 0.0292 - val_loss: 0.2184 - val_mse: 0.0994\n",
      "Epoch 121/200\n",
      "851/851 [==============================] - 14s 16ms/step - loss: 0.1270 - mse: 0.0287 - val_loss: 0.2160 - val_mse: 0.0978\n",
      "Epoch 122/200\n",
      "851/851 [==============================] - 13s 16ms/step - loss: 0.1287 - mse: 0.0288 - val_loss: 0.2043 - val_mse: 0.0914\n",
      "Epoch 123/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1275 - mse: 0.0285 - val_loss: 0.2198 - val_mse: 0.1013\n",
      "Epoch 124/200\n",
      "851/851 [==============================] - 12s 14ms/step - loss: 0.1273 - mse: 0.0285 - val_loss: 0.2079 - val_mse: 0.0933\n",
      "Epoch 125/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1256 - mse: 0.0279 - val_loss: 0.2133 - val_mse: 0.0935\n",
      "Epoch 126/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1265 - mse: 0.0285 - val_loss: 0.2092 - val_mse: 0.0953\n",
      "Epoch 127/200\n",
      "851/851 [==============================] - 12s 15ms/step - loss: 0.1266 - mse: 0.0284 - val_loss: 0.2040 - val_mse: 0.0885\n",
      "Epoch 128/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1261 - mse: 0.0284 - val_loss: 0.2141 - val_mse: 0.0990\n",
      "Epoch 129/200\n",
      "851/851 [==============================] - 13s 16ms/step - loss: 0.1245 - mse: 0.0276 - val_loss: 0.2164 - val_mse: 0.1042\n",
      "Epoch 130/200\n",
      "851/851 [==============================] - 14s 16ms/step - loss: 0.1249 - mse: 0.0272 - val_loss: 0.2162 - val_mse: 0.1011\n",
      "Epoch 131/200\n",
      "851/851 [==============================] - 13s 16ms/step - loss: 0.1248 - mse: 0.0278 - val_loss: 0.2075 - val_mse: 0.0937\n",
      "Epoch 132/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1249 - mse: 0.0278 - val_loss: 0.2026 - val_mse: 0.0868\n",
      "Epoch 133/200\n",
      "851/851 [==============================] - 13s 15ms/step - loss: 0.1246 - mse: 0.0277 - val_loss: 0.2171 - val_mse: 0.0982\n",
      "Epoch 134/200\n",
      "851/851 [==============================] - 14s 16ms/step - loss: 0.1243 - mse: 0.0275 - val_loss: 0.2046 - val_mse: 0.0910\n"
     ]
    }
   ],
   "source": [
    "X_train, X_val, X_test, y_train, y_val, y_test, scaler,yscaler= create_sequences_3(df, window_size=window_size_2, test_size=0.2)\n",
    "\n",
    "model_2 = Sequential([\n",
    "    GRU(64, return_sequences=True, input_shape=(window_size_2, num_features)),\n",
    "    Dropout(0.3),\n",
    "    GRU(32, return_sequences=True),\n",
    "    Dropout(0.2),\n",
    "    GRU(16, return_sequences=False),\n",
    "    Dropout(0.1),\n",
    "    Dense(32, activation='relu'),\n",
    "    Dense(1)  # Output layer for regression\n",
    "])\n",
    "model_2.compile(optimizer=Adam(learning_rate=0.001),loss='mae',metrics=['mse'])\n",
    "# Train the model\n",
    "history_2 = model_2.fit(X_train,y_train,\n",
    "                    validation_data=(X_val,y_val),\n",
    "                    epochs=200, batch_size=16,\n",
    "                    callbacks=tf.keras.callbacks.EarlyStopping(patience=30,monitor='val_loss',restore_best_weights=True))\n",
    "model_2.save('model_2.keras')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/200\n",
      "808/808 [==============================] - 18s 17ms/step - loss: 0.6161 - mse: 0.5903 - val_loss: 0.6216 - val_mse: 0.6089\n",
      "Epoch 2/200\n",
      "808/808 [==============================] - 13s 16ms/step - loss: 0.5933 - mse: 0.5593 - val_loss: 0.6024 - val_mse: 0.5586\n",
      "Epoch 3/200\n",
      "808/808 [==============================] - 15s 19ms/step - loss: 0.5645 - mse: 0.5184 - val_loss: 0.5526 - val_mse: 0.4960\n",
      "Epoch 4/200\n",
      "808/808 [==============================] - 14s 18ms/step - loss: 0.5317 - mse: 0.4754 - val_loss: 0.5137 - val_mse: 0.4357\n",
      "Epoch 5/200\n",
      "808/808 [==============================] - 15s 19ms/step - loss: 0.4835 - mse: 0.4040 - val_loss: 0.4754 - val_mse: 0.3907\n",
      "Epoch 6/200\n",
      "808/808 [==============================] - 15s 18ms/step - loss: 0.4316 - mse: 0.3366 - val_loss: 0.4146 - val_mse: 0.3036\n",
      "Epoch 7/200\n",
      "808/808 [==============================] - 15s 18ms/step - loss: 0.3807 - mse: 0.2734 - val_loss: 0.3953 - val_mse: 0.2695\n",
      "Epoch 8/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.3341 - mse: 0.2151 - val_loss: 0.3311 - val_mse: 0.1993\n",
      "Epoch 9/200\n",
      "808/808 [==============================] - 15s 18ms/step - loss: 0.2939 - mse: 0.1678 - val_loss: 0.2842 - val_mse: 0.1521\n",
      "Epoch 10/200\n",
      "808/808 [==============================] - 14s 18ms/step - loss: 0.2643 - mse: 0.1342 - val_loss: 0.2606 - val_mse: 0.1306\n",
      "Epoch 11/200\n",
      "808/808 [==============================] - 15s 18ms/step - loss: 0.2373 - mse: 0.1076 - val_loss: 0.2682 - val_mse: 0.1350\n",
      "Epoch 12/200\n",
      "808/808 [==============================] - 15s 18ms/step - loss: 0.2213 - mse: 0.0912 - val_loss: 0.2110 - val_mse: 0.0887\n",
      "Epoch 13/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.2028 - mse: 0.0770 - val_loss: 0.1861 - val_mse: 0.0696\n",
      "Epoch 14/200\n",
      "808/808 [==============================] - 15s 19ms/step - loss: 0.1921 - mse: 0.0682 - val_loss: 0.2135 - val_mse: 0.0814\n",
      "Epoch 15/200\n",
      "808/808 [==============================] - 15s 18ms/step - loss: 0.1803 - mse: 0.0595 - val_loss: 0.1711 - val_mse: 0.0557\n",
      "Epoch 16/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.1696 - mse: 0.0524 - val_loss: 0.1693 - val_mse: 0.0565\n",
      "Epoch 17/200\n",
      "808/808 [==============================] - 15s 18ms/step - loss: 0.1620 - mse: 0.0477 - val_loss: 0.1566 - val_mse: 0.0499\n",
      "Epoch 18/200\n",
      "808/808 [==============================] - 14s 18ms/step - loss: 0.1581 - mse: 0.0449 - val_loss: 0.1475 - val_mse: 0.0416\n",
      "Epoch 19/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.1489 - mse: 0.0400 - val_loss: 0.1364 - val_mse: 0.0348\n",
      "Epoch 20/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.1470 - mse: 0.0387 - val_loss: 0.1353 - val_mse: 0.0334\n",
      "Epoch 21/200\n",
      "808/808 [==============================] - 14s 18ms/step - loss: 0.1411 - mse: 0.0353 - val_loss: 0.1233 - val_mse: 0.0279\n",
      "Epoch 22/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.1401 - mse: 0.0344 - val_loss: 0.1412 - val_mse: 0.0358\n",
      "Epoch 23/200\n",
      "808/808 [==============================] - 14s 18ms/step - loss: 0.1347 - mse: 0.0318 - val_loss: 0.1211 - val_mse: 0.0276\n",
      "Epoch 24/200\n",
      "808/808 [==============================] - 15s 18ms/step - loss: 0.1288 - mse: 0.0291 - val_loss: 0.1156 - val_mse: 0.0257\n",
      "Epoch 25/200\n",
      "808/808 [==============================] - 14s 18ms/step - loss: 0.1265 - mse: 0.0280 - val_loss: 0.1253 - val_mse: 0.0289\n",
      "Epoch 26/200\n",
      "808/808 [==============================] - 15s 18ms/step - loss: 0.1246 - mse: 0.0274 - val_loss: 0.1426 - val_mse: 0.0348\n",
      "Epoch 27/200\n",
      "808/808 [==============================] - 15s 18ms/step - loss: 0.1235 - mse: 0.0267 - val_loss: 0.1087 - val_mse: 0.0218\n",
      "Epoch 28/200\n",
      "808/808 [==============================] - 15s 18ms/step - loss: 0.1182 - mse: 0.0247 - val_loss: 0.1129 - val_mse: 0.0234\n",
      "Epoch 29/200\n",
      "808/808 [==============================] - 16s 20ms/step - loss: 0.1183 - mse: 0.0247 - val_loss: 0.1120 - val_mse: 0.0223\n",
      "Epoch 30/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.1145 - mse: 0.0229 - val_loss: 0.1108 - val_mse: 0.0226\n",
      "Epoch 31/200\n",
      "808/808 [==============================] - 13s 16ms/step - loss: 0.1137 - mse: 0.0227 - val_loss: 0.1056 - val_mse: 0.0200\n",
      "Epoch 32/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.1112 - mse: 0.0219 - val_loss: 0.1081 - val_mse: 0.0211\n",
      "Epoch 33/200\n",
      "808/808 [==============================] - 16s 19ms/step - loss: 0.1101 - mse: 0.0210 - val_loss: 0.1069 - val_mse: 0.0217\n",
      "Epoch 34/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.1072 - mse: 0.0201 - val_loss: 0.0956 - val_mse: 0.0170\n",
      "Epoch 35/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.1067 - mse: 0.0200 - val_loss: 0.1242 - val_mse: 0.0276\n",
      "Epoch 36/200\n",
      "808/808 [==============================] - 14s 18ms/step - loss: 0.1048 - mse: 0.0191 - val_loss: 0.1047 - val_mse: 0.0201\n",
      "Epoch 37/200\n",
      "808/808 [==============================] - 15s 18ms/step - loss: 0.1040 - mse: 0.0191 - val_loss: 0.1064 - val_mse: 0.0205\n",
      "Epoch 38/200\n",
      "808/808 [==============================] - 15s 19ms/step - loss: 0.1029 - mse: 0.0187 - val_loss: 0.1006 - val_mse: 0.0193\n",
      "Epoch 39/200\n",
      "808/808 [==============================] - 16s 20ms/step - loss: 0.1003 - mse: 0.0176 - val_loss: 0.0967 - val_mse: 0.0164\n",
      "Epoch 40/200\n",
      "808/808 [==============================] - 15s 19ms/step - loss: 0.1004 - mse: 0.0178 - val_loss: 0.1218 - val_mse: 0.0254\n",
      "Epoch 41/200\n",
      "808/808 [==============================] - 14s 18ms/step - loss: 0.0987 - mse: 0.0171 - val_loss: 0.1104 - val_mse: 0.0217\n",
      "Epoch 42/200\n",
      "808/808 [==============================] - 15s 19ms/step - loss: 0.0975 - mse: 0.0166 - val_loss: 0.0874 - val_mse: 0.0136\n",
      "Epoch 43/200\n",
      "808/808 [==============================] - 15s 19ms/step - loss: 0.0978 - mse: 0.0167 - val_loss: 0.1062 - val_mse: 0.0205\n",
      "Epoch 44/200\n",
      "808/808 [==============================] - 13s 16ms/step - loss: 0.0965 - mse: 0.0162 - val_loss: 0.0981 - val_mse: 0.0174\n",
      "Epoch 45/200\n",
      "808/808 [==============================] - 13s 16ms/step - loss: 0.0950 - mse: 0.0160 - val_loss: 0.0960 - val_mse: 0.0162\n",
      "Epoch 46/200\n",
      "808/808 [==============================] - 13s 16ms/step - loss: 0.0940 - mse: 0.0153 - val_loss: 0.1222 - val_mse: 0.0264\n",
      "Epoch 47/200\n",
      "808/808 [==============================] - 13s 16ms/step - loss: 0.0951 - mse: 0.0156 - val_loss: 0.1007 - val_mse: 0.0187\n",
      "Epoch 48/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.0928 - mse: 0.0151 - val_loss: 0.1033 - val_mse: 0.0196\n",
      "Epoch 49/200\n",
      "808/808 [==============================] - 14s 18ms/step - loss: 0.0920 - mse: 0.0149 - val_loss: 0.0978 - val_mse: 0.0169\n",
      "Epoch 50/200\n",
      "808/808 [==============================] - 16s 19ms/step - loss: 0.0908 - mse: 0.0145 - val_loss: 0.1069 - val_mse: 0.0202\n",
      "Epoch 51/200\n",
      "808/808 [==============================] - 14s 18ms/step - loss: 0.0901 - mse: 0.0144 - val_loss: 0.1130 - val_mse: 0.0212\n",
      "Epoch 52/200\n",
      "808/808 [==============================] - 13s 17ms/step - loss: 0.0891 - mse: 0.0139 - val_loss: 0.1102 - val_mse: 0.0212\n",
      "Epoch 53/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.0879 - mse: 0.0134 - val_loss: 0.0871 - val_mse: 0.0139\n",
      "Epoch 54/200\n",
      "808/808 [==============================] - 14s 18ms/step - loss: 0.0881 - mse: 0.0137 - val_loss: 0.0844 - val_mse: 0.0138\n",
      "Epoch 55/200\n",
      "808/808 [==============================] - 15s 18ms/step - loss: 0.0877 - mse: 0.0135 - val_loss: 0.1058 - val_mse: 0.0205\n",
      "Epoch 56/200\n",
      "808/808 [==============================] - 14s 18ms/step - loss: 0.0860 - mse: 0.0129 - val_loss: 0.1087 - val_mse: 0.0207\n",
      "Epoch 57/200\n",
      "808/808 [==============================] - 15s 18ms/step - loss: 0.0856 - mse: 0.0127 - val_loss: 0.0972 - val_mse: 0.0171\n",
      "Epoch 58/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.0848 - mse: 0.0125 - val_loss: 0.0975 - val_mse: 0.0163\n",
      "Epoch 59/200\n",
      "808/808 [==============================] - 13s 17ms/step - loss: 0.0840 - mse: 0.0125 - val_loss: 0.0945 - val_mse: 0.0151\n",
      "Epoch 60/200\n",
      "808/808 [==============================] - 13s 16ms/step - loss: 0.0819 - mse: 0.0120 - val_loss: 0.1047 - val_mse: 0.0200\n",
      "Epoch 61/200\n",
      "808/808 [==============================] - 13s 16ms/step - loss: 0.0834 - mse: 0.0120 - val_loss: 0.0920 - val_mse: 0.0152\n",
      "Epoch 62/200\n",
      "808/808 [==============================] - 13s 16ms/step - loss: 0.0822 - mse: 0.0118 - val_loss: 0.0979 - val_mse: 0.0174\n",
      "Epoch 63/200\n",
      "808/808 [==============================] - 13s 16ms/step - loss: 0.0817 - mse: 0.0118 - val_loss: 0.1021 - val_mse: 0.0184\n",
      "Epoch 64/200\n",
      "808/808 [==============================] - 13s 16ms/step - loss: 0.0821 - mse: 0.0118 - val_loss: 0.0935 - val_mse: 0.0164\n",
      "Epoch 65/200\n",
      "808/808 [==============================] - 13s 16ms/step - loss: 0.0809 - mse: 0.0115 - val_loss: 0.0976 - val_mse: 0.0171\n",
      "Epoch 66/200\n",
      "808/808 [==============================] - 13s 16ms/step - loss: 0.0797 - mse: 0.0111 - val_loss: 0.1005 - val_mse: 0.0181\n",
      "Epoch 67/200\n",
      "808/808 [==============================] - 13s 16ms/step - loss: 0.0804 - mse: 0.0114 - val_loss: 0.0908 - val_mse: 0.0148\n",
      "Epoch 68/200\n",
      "808/808 [==============================] - 13s 16ms/step - loss: 0.0799 - mse: 0.0112 - val_loss: 0.1018 - val_mse: 0.0191\n",
      "Epoch 69/200\n",
      "808/808 [==============================] - 13s 16ms/step - loss: 0.0790 - mse: 0.0111 - val_loss: 0.0867 - val_mse: 0.0140\n",
      "Epoch 70/200\n",
      "808/808 [==============================] - 13s 16ms/step - loss: 0.0787 - mse: 0.0108 - val_loss: 0.1052 - val_mse: 0.0194\n",
      "Epoch 71/200\n",
      "808/808 [==============================] - 13s 16ms/step - loss: 0.0787 - mse: 0.0108 - val_loss: 0.0956 - val_mse: 0.0166\n",
      "Epoch 72/200\n",
      "808/808 [==============================] - 13s 16ms/step - loss: 0.0783 - mse: 0.0107 - val_loss: 0.0921 - val_mse: 0.0158\n",
      "Epoch 73/200\n",
      "808/808 [==============================] - 13s 16ms/step - loss: 0.0759 - mse: 0.0100 - val_loss: 0.0996 - val_mse: 0.0170\n",
      "Epoch 74/200\n",
      "808/808 [==============================] - 13s 16ms/step - loss: 0.0772 - mse: 0.0104 - val_loss: 0.0990 - val_mse: 0.0185\n",
      "Epoch 75/200\n",
      "808/808 [==============================] - 13s 16ms/step - loss: 0.0762 - mse: 0.0103 - val_loss: 0.0830 - val_mse: 0.0128\n",
      "Epoch 76/200\n",
      "808/808 [==============================] - 13s 16ms/step - loss: 0.0761 - mse: 0.0100 - val_loss: 0.0890 - val_mse: 0.0138\n",
      "Epoch 77/200\n",
      "808/808 [==============================] - 13s 17ms/step - loss: 0.0755 - mse: 0.0098 - val_loss: 0.0891 - val_mse: 0.0138\n",
      "Epoch 78/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.0758 - mse: 0.0100 - val_loss: 0.1062 - val_mse: 0.0203\n",
      "Epoch 79/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.0740 - mse: 0.0098 - val_loss: 0.0856 - val_mse: 0.0136\n",
      "Epoch 80/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.0734 - mse: 0.0095 - val_loss: 0.1012 - val_mse: 0.0183\n",
      "Epoch 81/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.0743 - mse: 0.0098 - val_loss: 0.0885 - val_mse: 0.0133\n",
      "Epoch 82/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.0743 - mse: 0.0096 - val_loss: 0.1047 - val_mse: 0.0195\n",
      "Epoch 83/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.0727 - mse: 0.0093 - val_loss: 0.0942 - val_mse: 0.0150\n",
      "Epoch 84/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.0731 - mse: 0.0096 - val_loss: 0.0894 - val_mse: 0.0137\n",
      "Epoch 85/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.0719 - mse: 0.0092 - val_loss: 0.1026 - val_mse: 0.0182\n",
      "Epoch 86/200\n",
      "808/808 [==============================] - 14s 18ms/step - loss: 0.0723 - mse: 0.0093 - val_loss: 0.1066 - val_mse: 0.0194\n",
      "Epoch 87/200\n",
      "808/808 [==============================] - 14s 18ms/step - loss: 0.0729 - mse: 0.0093 - val_loss: 0.0921 - val_mse: 0.0150\n",
      "Epoch 88/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.0714 - mse: 0.0089 - val_loss: 0.0898 - val_mse: 0.0143\n",
      "Epoch 89/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.0722 - mse: 0.0093 - val_loss: 0.1116 - val_mse: 0.0227\n",
      "Epoch 90/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.0722 - mse: 0.0093 - val_loss: 0.0980 - val_mse: 0.0170\n",
      "Epoch 91/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.0709 - mse: 0.0089 - val_loss: 0.1166 - val_mse: 0.0230\n",
      "Epoch 92/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.0711 - mse: 0.0088 - val_loss: 0.0814 - val_mse: 0.0121\n",
      "Epoch 93/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.0709 - mse: 0.0091 - val_loss: 0.0816 - val_mse: 0.0127\n",
      "Epoch 94/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.0693 - mse: 0.0086 - val_loss: 0.0753 - val_mse: 0.0102\n",
      "Epoch 95/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.0707 - mse: 0.0089 - val_loss: 0.1155 - val_mse: 0.0244\n",
      "Epoch 96/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.0693 - mse: 0.0085 - val_loss: 0.0956 - val_mse: 0.0178\n",
      "Epoch 97/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.0694 - mse: 0.0086 - val_loss: 0.1017 - val_mse: 0.0182\n",
      "Epoch 98/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.0704 - mse: 0.0088 - val_loss: 0.1017 - val_mse: 0.0181\n",
      "Epoch 99/200\n",
      "808/808 [==============================] - 14s 18ms/step - loss: 0.0693 - mse: 0.0084 - val_loss: 0.0925 - val_mse: 0.0156\n",
      "Epoch 100/200\n",
      "808/808 [==============================] - 15s 18ms/step - loss: 0.0694 - mse: 0.0089 - val_loss: 0.0913 - val_mse: 0.0147\n",
      "Epoch 101/200\n",
      "808/808 [==============================] - 14s 18ms/step - loss: 0.0689 - mse: 0.0085 - val_loss: 0.0998 - val_mse: 0.0188\n",
      "Epoch 102/200\n",
      "808/808 [==============================] - 13s 16ms/step - loss: 0.0676 - mse: 0.0081 - val_loss: 0.0998 - val_mse: 0.0178\n",
      "Epoch 103/200\n",
      "808/808 [==============================] - 14s 18ms/step - loss: 0.0670 - mse: 0.0079 - val_loss: 0.0966 - val_mse: 0.0171\n",
      "Epoch 104/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.0675 - mse: 0.0081 - val_loss: 0.0835 - val_mse: 0.0133\n",
      "Epoch 105/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.0673 - mse: 0.0080 - val_loss: 0.1015 - val_mse: 0.0183\n",
      "Epoch 106/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.0680 - mse: 0.0084 - val_loss: 0.1090 - val_mse: 0.0202\n",
      "Epoch 107/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.0663 - mse: 0.0077 - val_loss: 0.1045 - val_mse: 0.0200\n",
      "Epoch 108/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.0662 - mse: 0.0077 - val_loss: 0.1121 - val_mse: 0.0227\n",
      "Epoch 109/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.0664 - mse: 0.0079 - val_loss: 0.0993 - val_mse: 0.0168\n",
      "Epoch 110/200\n",
      "808/808 [==============================] - 15s 18ms/step - loss: 0.0659 - mse: 0.0078 - val_loss: 0.1112 - val_mse: 0.0214\n",
      "Epoch 111/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.0659 - mse: 0.0079 - val_loss: 0.0967 - val_mse: 0.0169\n",
      "Epoch 112/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.0665 - mse: 0.0080 - val_loss: 0.0890 - val_mse: 0.0149\n",
      "Epoch 113/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.0649 - mse: 0.0074 - val_loss: 0.0849 - val_mse: 0.0130\n",
      "Epoch 114/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.0662 - mse: 0.0077 - val_loss: 0.0926 - val_mse: 0.0156\n",
      "Epoch 115/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.0648 - mse: 0.0075 - val_loss: 0.0918 - val_mse: 0.0152\n",
      "Epoch 116/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.0647 - mse: 0.0073 - val_loss: 0.1136 - val_mse: 0.0214\n",
      "Epoch 117/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.0647 - mse: 0.0074 - val_loss: 0.1084 - val_mse: 0.0210\n",
      "Epoch 118/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.0650 - mse: 0.0076 - val_loss: 0.0888 - val_mse: 0.0148\n",
      "Epoch 119/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.0644 - mse: 0.0074 - val_loss: 0.0977 - val_mse: 0.0176\n",
      "Epoch 120/200\n",
      "808/808 [==============================] - 14s 17ms/step - loss: 0.0641 - mse: 0.0072 - val_loss: 0.0922 - val_mse: 0.0165\n",
      "Epoch 121/200\n",
      "808/808 [==============================] - 14s 18ms/step - loss: 0.0636 - mse: 0.0072 - val_loss: 0.0970 - val_mse: 0.0174\n",
      "Epoch 122/200\n",
      "808/808 [==============================] - 15s 19ms/step - loss: 0.0640 - mse: 0.0074 - val_loss: 0.0860 - val_mse: 0.0137\n",
      "Epoch 123/200\n",
      "808/808 [==============================] - 16s 20ms/step - loss: 0.0640 - mse: 0.0073 - val_loss: 0.1083 - val_mse: 0.0209\n",
      "Epoch 124/200\n",
      "808/808 [==============================] - 16s 20ms/step - loss: 0.0633 - mse: 0.0071 - val_loss: 0.0987 - val_mse: 0.0194\n"
     ]
    }
   ],
   "source": [
    "X_train, X_val, X_test, y_train, y_val, y_test, scaler,yscaler= create_sequences_3(df, window_size=window_size_3, test_size=0.2)\n",
    "\n",
    "model_3 = Sequential([\n",
    "    GRU(64, return_sequences=True, input_shape=(window_size_3, num_features)),\n",
    "    Dropout(0.3),\n",
    "    GRU(32, return_sequences=True),\n",
    "    Dropout(0.2),\n",
    "    GRU(16, return_sequences=False),\n",
    "    Dropout(0.1),\n",
    "    Dense(32, activation='relu'),\n",
    "    Dense(1)  # Output layer for regression\n",
    "])\n",
    "model_3.compile(optimizer=Adam(learning_rate=0.001),loss='mae',metrics=['mse'])\n",
    "# Train the model\n",
    "history_3 = model_3.fit(X_train,y_train,\n",
    "                    validation_data=(X_val,y_val),\n",
    "                    epochs=200, batch_size=16,\n",
    "                    callbacks=tf.keras.callbacks.EarlyStopping(patience=30,monitor='val_loss',restore_best_weights=True))\n",
    "model_3.save('model_3.keras')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/200\n",
      "764/764 [==============================] - 26s 27ms/step - loss: 0.6214 - mse: 0.5973 - val_loss: 0.6134 - val_mse: 0.6086\n",
      "Epoch 2/200\n",
      "764/764 [==============================] - 19s 25ms/step - loss: 0.5999 - mse: 0.5670 - val_loss: 0.5602 - val_mse: 0.5065\n",
      "Epoch 3/200\n",
      "764/764 [==============================] - 20s 26ms/step - loss: 0.5524 - mse: 0.5012 - val_loss: 0.5295 - val_mse: 0.4801\n",
      "Epoch 4/200\n",
      "764/764 [==============================] - 19s 25ms/step - loss: 0.4860 - mse: 0.4073 - val_loss: 0.4302 - val_mse: 0.3294\n",
      "Epoch 5/200\n",
      "764/764 [==============================] - 19s 25ms/step - loss: 0.4183 - mse: 0.3179 - val_loss: 0.3625 - val_mse: 0.2496\n",
      "Epoch 6/200\n",
      "764/764 [==============================] - 19s 25ms/step - loss: 0.3457 - mse: 0.2245 - val_loss: 0.3060 - val_mse: 0.1745\n",
      "Epoch 7/200\n",
      "764/764 [==============================] - 19s 25ms/step - loss: 0.2865 - mse: 0.1548 - val_loss: 0.2453 - val_mse: 0.1094\n",
      "Epoch 8/200\n",
      "764/764 [==============================] - 18s 24ms/step - loss: 0.2502 - mse: 0.1177 - val_loss: 0.2320 - val_mse: 0.0963\n",
      "Epoch 9/200\n",
      "764/764 [==============================] - 18s 24ms/step - loss: 0.2171 - mse: 0.0874 - val_loss: 0.1760 - val_mse: 0.0579\n",
      "Epoch 10/200\n",
      "764/764 [==============================] - 19s 25ms/step - loss: 0.1918 - mse: 0.0685 - val_loss: 0.1622 - val_mse: 0.0477\n",
      "Epoch 11/200\n",
      "764/764 [==============================] - 19s 24ms/step - loss: 0.1782 - mse: 0.0580 - val_loss: 0.1429 - val_mse: 0.0379\n",
      "Epoch 12/200\n",
      "764/764 [==============================] - 19s 25ms/step - loss: 0.1648 - mse: 0.0492 - val_loss: 0.1432 - val_mse: 0.0363\n",
      "Epoch 13/200\n",
      "764/764 [==============================] - 19s 25ms/step - loss: 0.1526 - mse: 0.0416 - val_loss: 0.1293 - val_mse: 0.0312\n",
      "Epoch 14/200\n",
      "764/764 [==============================] - 19s 25ms/step - loss: 0.1455 - mse: 0.0374 - val_loss: 0.1146 - val_mse: 0.0245\n",
      "Epoch 15/200\n",
      "764/764 [==============================] - 19s 25ms/step - loss: 0.1382 - mse: 0.0337 - val_loss: 0.1207 - val_mse: 0.0267\n",
      "Epoch 16/200\n",
      "764/764 [==============================] - 18s 23ms/step - loss: 0.1309 - mse: 0.0302 - val_loss: 0.1117 - val_mse: 0.0218\n",
      "Epoch 17/200\n",
      "764/764 [==============================] - 19s 24ms/step - loss: 0.1269 - mse: 0.0286 - val_loss: 0.1098 - val_mse: 0.0216\n",
      "Epoch 18/200\n",
      "764/764 [==============================] - 19s 25ms/step - loss: 0.1226 - mse: 0.0269 - val_loss: 0.1172 - val_mse: 0.0246\n",
      "Epoch 19/200\n",
      "764/764 [==============================] - 19s 25ms/step - loss: 0.1214 - mse: 0.0261 - val_loss: 0.1144 - val_mse: 0.0229\n",
      "Epoch 20/200\n",
      "764/764 [==============================] - 19s 25ms/step - loss: 0.1175 - mse: 0.0246 - val_loss: 0.0932 - val_mse: 0.0154\n",
      "Epoch 21/200\n",
      "764/764 [==============================] - 19s 25ms/step - loss: 0.1148 - mse: 0.0231 - val_loss: 0.1016 - val_mse: 0.0177\n",
      "Epoch 22/200\n",
      "764/764 [==============================] - 18s 23ms/step - loss: 0.1102 - mse: 0.0213 - val_loss: 0.1148 - val_mse: 0.0222\n",
      "Epoch 23/200\n",
      "764/764 [==============================] - 20s 26ms/step - loss: 0.1084 - mse: 0.0203 - val_loss: 0.0925 - val_mse: 0.0154\n",
      "Epoch 24/200\n",
      "764/764 [==============================] - 19s 25ms/step - loss: 0.1076 - mse: 0.0204 - val_loss: 0.0989 - val_mse: 0.0175\n",
      "Epoch 25/200\n",
      "764/764 [==============================] - 20s 26ms/step - loss: 0.1038 - mse: 0.0189 - val_loss: 0.0912 - val_mse: 0.0139\n",
      "Epoch 26/200\n",
      "764/764 [==============================] - 19s 25ms/step - loss: 0.1036 - mse: 0.0189 - val_loss: 0.0918 - val_mse: 0.0150\n",
      "Epoch 27/200\n",
      "764/764 [==============================] - 19s 25ms/step - loss: 0.1001 - mse: 0.0177 - val_loss: 0.0960 - val_mse: 0.0163\n",
      "Epoch 28/200\n",
      "764/764 [==============================] - 17s 23ms/step - loss: 0.0980 - mse: 0.0168 - val_loss: 0.0918 - val_mse: 0.0138\n",
      "Epoch 29/200\n",
      "764/764 [==============================] - 17s 23ms/step - loss: 0.0954 - mse: 0.0158 - val_loss: 0.1154 - val_mse: 0.0211\n",
      "Epoch 30/200\n",
      "764/764 [==============================] - 18s 24ms/step - loss: 0.0930 - mse: 0.0152 - val_loss: 0.0872 - val_mse: 0.0131\n",
      "Epoch 31/200\n",
      "764/764 [==============================] - 19s 25ms/step - loss: 0.0934 - mse: 0.0153 - val_loss: 0.0885 - val_mse: 0.0135\n",
      "Epoch 32/200\n",
      "764/764 [==============================] - 19s 25ms/step - loss: 0.0925 - mse: 0.0153 - val_loss: 0.0866 - val_mse: 0.0123\n",
      "Epoch 33/200\n",
      "764/764 [==============================] - 18s 24ms/step - loss: 0.0899 - mse: 0.0141 - val_loss: 0.1043 - val_mse: 0.0192\n",
      "Epoch 34/200\n",
      "764/764 [==============================] - 18s 23ms/step - loss: 0.0887 - mse: 0.0137 - val_loss: 0.0855 - val_mse: 0.0127\n",
      "Epoch 35/200\n",
      "764/764 [==============================] - 18s 24ms/step - loss: 0.0870 - mse: 0.0132 - val_loss: 0.0783 - val_mse: 0.0108\n",
      "Epoch 36/200\n",
      "764/764 [==============================] - 19s 25ms/step - loss: 0.0872 - mse: 0.0134 - val_loss: 0.1037 - val_mse: 0.0178\n",
      "Epoch 37/200\n",
      "764/764 [==============================] - 19s 25ms/step - loss: 0.0841 - mse: 0.0123 - val_loss: 0.1189 - val_mse: 0.0227\n",
      "Epoch 38/200\n",
      "764/764 [==============================] - 19s 25ms/step - loss: 0.0835 - mse: 0.0122 - val_loss: 0.0806 - val_mse: 0.0110\n",
      "Epoch 39/200\n",
      "764/764 [==============================] - 19s 25ms/step - loss: 0.0846 - mse: 0.0125 - val_loss: 0.0975 - val_mse: 0.0153\n",
      "Epoch 40/200\n",
      "764/764 [==============================] - 19s 24ms/step - loss: 0.0808 - mse: 0.0114 - val_loss: 0.0922 - val_mse: 0.0141\n",
      "Epoch 41/200\n",
      "764/764 [==============================] - 19s 25ms/step - loss: 0.0812 - mse: 0.0115 - val_loss: 0.0797 - val_mse: 0.0102\n",
      "Epoch 42/200\n",
      "764/764 [==============================] - 19s 24ms/step - loss: 0.0797 - mse: 0.0111 - val_loss: 0.0932 - val_mse: 0.0151\n",
      "Epoch 43/200\n",
      "764/764 [==============================] - 18s 23ms/step - loss: 0.0794 - mse: 0.0111 - val_loss: 0.0942 - val_mse: 0.0146\n",
      "Epoch 44/200\n",
      "764/764 [==============================] - 18s 23ms/step - loss: 0.0782 - mse: 0.0105 - val_loss: 0.0791 - val_mse: 0.0101\n",
      "Epoch 45/200\n",
      "764/764 [==============================] - 18s 24ms/step - loss: 0.0771 - mse: 0.0103 - val_loss: 0.0918 - val_mse: 0.0145\n",
      "Epoch 46/200\n",
      "764/764 [==============================] - 15s 19ms/step - loss: 0.0767 - mse: 0.0103 - val_loss: 0.0896 - val_mse: 0.0131\n",
      "Epoch 47/200\n",
      "764/764 [==============================] - 17s 22ms/step - loss: 0.0756 - mse: 0.0101 - val_loss: 0.0826 - val_mse: 0.0114\n",
      "Epoch 48/200\n",
      "764/764 [==============================] - 15s 20ms/step - loss: 0.0767 - mse: 0.0103 - val_loss: 0.0948 - val_mse: 0.0138\n",
      "Epoch 49/200\n",
      "764/764 [==============================] - 16s 21ms/step - loss: 0.0737 - mse: 0.0095 - val_loss: 0.0913 - val_mse: 0.0131\n",
      "Epoch 50/200\n",
      "764/764 [==============================] - 17s 22ms/step - loss: 0.0740 - mse: 0.0097 - val_loss: 0.0812 - val_mse: 0.0110\n",
      "Epoch 51/200\n",
      "764/764 [==============================] - 15s 20ms/step - loss: 0.0730 - mse: 0.0095 - val_loss: 0.0803 - val_mse: 0.0107\n",
      "Epoch 52/200\n",
      "764/764 [==============================] - 15s 20ms/step - loss: 0.0721 - mse: 0.0091 - val_loss: 0.0962 - val_mse: 0.0152\n",
      "Epoch 53/200\n",
      "764/764 [==============================] - 15s 20ms/step - loss: 0.0722 - mse: 0.0091 - val_loss: 0.0851 - val_mse: 0.0121\n",
      "Epoch 54/200\n",
      "764/764 [==============================] - 17s 22ms/step - loss: 0.0703 - mse: 0.0088 - val_loss: 0.0766 - val_mse: 0.0103\n",
      "Epoch 55/200\n",
      "764/764 [==============================] - 16s 21ms/step - loss: 0.0707 - mse: 0.0087 - val_loss: 0.0759 - val_mse: 0.0097\n",
      "Epoch 56/200\n",
      "764/764 [==============================] - 15s 20ms/step - loss: 0.0712 - mse: 0.0088 - val_loss: 0.0844 - val_mse: 0.0111\n",
      "Epoch 57/200\n",
      "764/764 [==============================] - 15s 20ms/step - loss: 0.0691 - mse: 0.0083 - val_loss: 0.1128 - val_mse: 0.0228\n",
      "Epoch 58/200\n",
      "764/764 [==============================] - 15s 19ms/step - loss: 0.0691 - mse: 0.0083 - val_loss: 0.0853 - val_mse: 0.0119\n",
      "Epoch 59/200\n",
      "764/764 [==============================] - 18s 24ms/step - loss: 0.0685 - mse: 0.0083 - val_loss: 0.0804 - val_mse: 0.0110\n",
      "Epoch 60/200\n",
      "764/764 [==============================] - 15s 20ms/step - loss: 0.0673 - mse: 0.0079 - val_loss: 0.0841 - val_mse: 0.0130\n",
      "Epoch 61/200\n",
      "764/764 [==============================] - 16s 21ms/step - loss: 0.0679 - mse: 0.0079 - val_loss: 0.0999 - val_mse: 0.0162\n",
      "Epoch 62/200\n",
      "764/764 [==============================] - 16s 21ms/step - loss: 0.0661 - mse: 0.0077 - val_loss: 0.0843 - val_mse: 0.0115\n",
      "Epoch 63/200\n",
      "764/764 [==============================] - 15s 19ms/step - loss: 0.0674 - mse: 0.0079 - val_loss: 0.0925 - val_mse: 0.0129\n",
      "Epoch 64/200\n",
      "764/764 [==============================] - 15s 19ms/step - loss: 0.0660 - mse: 0.0076 - val_loss: 0.0981 - val_mse: 0.0155\n",
      "Epoch 65/200\n",
      "764/764 [==============================] - 16s 21ms/step - loss: 0.0661 - mse: 0.0077 - val_loss: 0.0723 - val_mse: 0.0084\n",
      "Epoch 66/200\n",
      "764/764 [==============================] - 16s 21ms/step - loss: 0.0653 - mse: 0.0074 - val_loss: 0.0843 - val_mse: 0.0112\n",
      "Epoch 67/200\n",
      "764/764 [==============================] - 15s 19ms/step - loss: 0.0648 - mse: 0.0073 - val_loss: 0.0810 - val_mse: 0.0110\n",
      "Epoch 68/200\n",
      "764/764 [==============================] - 15s 20ms/step - loss: 0.0642 - mse: 0.0073 - val_loss: 0.0778 - val_mse: 0.0106\n",
      "Epoch 69/200\n",
      "764/764 [==============================] - 15s 20ms/step - loss: 0.0639 - mse: 0.0071 - val_loss: 0.0918 - val_mse: 0.0141\n",
      "Epoch 70/200\n",
      "764/764 [==============================] - 16s 21ms/step - loss: 0.0628 - mse: 0.0071 - val_loss: 0.0751 - val_mse: 0.0098\n",
      "Epoch 71/200\n",
      "764/764 [==============================] - 15s 19ms/step - loss: 0.0633 - mse: 0.0069 - val_loss: 0.0935 - val_mse: 0.0131\n",
      "Epoch 72/200\n",
      "764/764 [==============================] - 16s 21ms/step - loss: 0.0619 - mse: 0.0068 - val_loss: 0.0819 - val_mse: 0.0121\n",
      "Epoch 73/200\n",
      "764/764 [==============================] - 17s 22ms/step - loss: 0.0624 - mse: 0.0068 - val_loss: 0.0899 - val_mse: 0.0131\n",
      "Epoch 74/200\n",
      "764/764 [==============================] - 18s 23ms/step - loss: 0.0625 - mse: 0.0070 - val_loss: 0.0987 - val_mse: 0.0155\n",
      "Epoch 75/200\n",
      "764/764 [==============================] - 16s 21ms/step - loss: 0.0622 - mse: 0.0067 - val_loss: 0.1012 - val_mse: 0.0162\n",
      "Epoch 76/200\n",
      "764/764 [==============================] - 15s 20ms/step - loss: 0.0608 - mse: 0.0065 - val_loss: 0.0919 - val_mse: 0.0141\n",
      "Epoch 77/200\n",
      "764/764 [==============================] - 14s 19ms/step - loss: 0.0612 - mse: 0.0067 - val_loss: 0.1087 - val_mse: 0.0197\n",
      "Epoch 78/200\n",
      "764/764 [==============================] - 16s 21ms/step - loss: 0.0601 - mse: 0.0063 - val_loss: 0.0754 - val_mse: 0.0094\n",
      "Epoch 79/200\n",
      "764/764 [==============================] - 14s 19ms/step - loss: 0.0593 - mse: 0.0062 - val_loss: 0.0822 - val_mse: 0.0118\n",
      "Epoch 80/200\n",
      "764/764 [==============================] - 14s 19ms/step - loss: 0.0592 - mse: 0.0061 - val_loss: 0.0951 - val_mse: 0.0138\n",
      "Epoch 81/200\n",
      "764/764 [==============================] - 14s 19ms/step - loss: 0.0595 - mse: 0.0062 - val_loss: 0.0877 - val_mse: 0.0122\n",
      "Epoch 82/200\n",
      "764/764 [==============================] - 15s 20ms/step - loss: 0.0600 - mse: 0.0063 - val_loss: 0.0730 - val_mse: 0.0092\n",
      "Epoch 83/200\n",
      "764/764 [==============================] - 16s 21ms/step - loss: 0.0594 - mse: 0.0061 - val_loss: 0.0985 - val_mse: 0.0169\n",
      "Epoch 84/200\n",
      "764/764 [==============================] - 16s 21ms/step - loss: 0.0595 - mse: 0.0062 - val_loss: 0.0843 - val_mse: 0.0119\n",
      "Epoch 85/200\n",
      "764/764 [==============================] - 17s 22ms/step - loss: 0.0577 - mse: 0.0059 - val_loss: 0.0815 - val_mse: 0.0109\n",
      "Epoch 86/200\n",
      "764/764 [==============================] - 14s 19ms/step - loss: 0.0588 - mse: 0.0063 - val_loss: 0.0772 - val_mse: 0.0095\n",
      "Epoch 87/200\n",
      "764/764 [==============================] - 14s 18ms/step - loss: 0.0582 - mse: 0.0060 - val_loss: 0.0975 - val_mse: 0.0144\n",
      "Epoch 88/200\n",
      "764/764 [==============================] - 15s 19ms/step - loss: 0.0582 - mse: 0.0059 - val_loss: 0.0805 - val_mse: 0.0108\n",
      "Epoch 89/200\n",
      "764/764 [==============================] - 15s 20ms/step - loss: 0.0568 - mse: 0.0056 - val_loss: 0.0945 - val_mse: 0.0139\n",
      "Epoch 90/200\n",
      "764/764 [==============================] - 15s 20ms/step - loss: 0.0587 - mse: 0.0061 - val_loss: 0.0843 - val_mse: 0.0108\n",
      "Epoch 91/200\n",
      "764/764 [==============================] - 15s 20ms/step - loss: 0.0574 - mse: 0.0058 - val_loss: 0.0992 - val_mse: 0.0153\n",
      "Epoch 92/200\n",
      "764/764 [==============================] - 14s 19ms/step - loss: 0.0563 - mse: 0.0057 - val_loss: 0.0662 - val_mse: 0.0073\n",
      "Epoch 93/200\n",
      "764/764 [==============================] - 15s 20ms/step - loss: 0.0563 - mse: 0.0056 - val_loss: 0.0969 - val_mse: 0.0155\n",
      "Epoch 94/200\n",
      "764/764 [==============================] - 14s 19ms/step - loss: 0.0552 - mse: 0.0054 - val_loss: 0.0752 - val_mse: 0.0100\n",
      "Epoch 95/200\n",
      "764/764 [==============================] - 15s 20ms/step - loss: 0.0560 - mse: 0.0056 - val_loss: 0.0855 - val_mse: 0.0125\n",
      "Epoch 96/200\n",
      "764/764 [==============================] - 15s 20ms/step - loss: 0.0566 - mse: 0.0059 - val_loss: 0.0882 - val_mse: 0.0118\n",
      "Epoch 97/200\n",
      "764/764 [==============================] - 15s 20ms/step - loss: 0.0550 - mse: 0.0053 - val_loss: 0.0858 - val_mse: 0.0125\n",
      "Epoch 98/200\n",
      "764/764 [==============================] - 15s 19ms/step - loss: 0.0542 - mse: 0.0052 - val_loss: 0.0906 - val_mse: 0.0133\n",
      "Epoch 99/200\n",
      "764/764 [==============================] - 15s 20ms/step - loss: 0.0550 - mse: 0.0054 - val_loss: 0.0898 - val_mse: 0.0128\n",
      "Epoch 100/200\n",
      "764/764 [==============================] - 15s 19ms/step - loss: 0.0539 - mse: 0.0052 - val_loss: 0.0954 - val_mse: 0.0151\n",
      "Epoch 101/200\n",
      "764/764 [==============================] - 15s 20ms/step - loss: 0.0544 - mse: 0.0052 - val_loss: 0.0665 - val_mse: 0.0083\n",
      "Epoch 102/200\n",
      "764/764 [==============================] - 15s 20ms/step - loss: 0.0545 - mse: 0.0054 - val_loss: 0.0778 - val_mse: 0.0097\n",
      "Epoch 103/200\n",
      "764/764 [==============================] - 15s 19ms/step - loss: 0.0545 - mse: 0.0053 - val_loss: 0.0900 - val_mse: 0.0124\n",
      "Epoch 104/200\n",
      "764/764 [==============================] - 14s 19ms/step - loss: 0.0538 - mse: 0.0050 - val_loss: 0.0912 - val_mse: 0.0132\n",
      "Epoch 105/200\n",
      "764/764 [==============================] - 15s 20ms/step - loss: 0.0541 - mse: 0.0052 - val_loss: 0.0797 - val_mse: 0.0100\n",
      "Epoch 106/200\n",
      "764/764 [==============================] - 15s 20ms/step - loss: 0.0528 - mse: 0.0050 - val_loss: 0.0743 - val_mse: 0.0092\n",
      "Epoch 107/200\n",
      "764/764 [==============================] - 15s 20ms/step - loss: 0.0531 - mse: 0.0050 - val_loss: 0.1067 - val_mse: 0.0178\n",
      "Epoch 108/200\n",
      "764/764 [==============================] - 16s 20ms/step - loss: 0.0527 - mse: 0.0049 - val_loss: 0.0844 - val_mse: 0.0123\n",
      "Epoch 109/200\n",
      "764/764 [==============================] - 15s 20ms/step - loss: 0.0523 - mse: 0.0049 - val_loss: 0.0848 - val_mse: 0.0123\n",
      "Epoch 110/200\n",
      "764/764 [==============================] - 15s 20ms/step - loss: 0.0525 - mse: 0.0048 - val_loss: 0.0802 - val_mse: 0.0099\n",
      "Epoch 111/200\n",
      "764/764 [==============================] - 15s 19ms/step - loss: 0.0521 - mse: 0.0047 - val_loss: 0.0887 - val_mse: 0.0126\n",
      "Epoch 112/200\n",
      "764/764 [==============================] - 15s 20ms/step - loss: 0.0522 - mse: 0.0049 - val_loss: 0.0982 - val_mse: 0.0145\n",
      "Epoch 113/200\n",
      "764/764 [==============================] - 14s 19ms/step - loss: 0.0524 - mse: 0.0049 - val_loss: 0.0828 - val_mse: 0.0114\n",
      "Epoch 114/200\n",
      "764/764 [==============================] - 15s 20ms/step - loss: 0.0515 - mse: 0.0048 - val_loss: 0.0917 - val_mse: 0.0133\n",
      "Epoch 115/200\n",
      "764/764 [==============================] - 15s 20ms/step - loss: 0.0512 - mse: 0.0047 - val_loss: 0.0736 - val_mse: 0.0097\n",
      "Epoch 116/200\n",
      "764/764 [==============================] - 15s 20ms/step - loss: 0.0519 - mse: 0.0049 - val_loss: 0.0866 - val_mse: 0.0122\n",
      "Epoch 117/200\n",
      "764/764 [==============================] - 16s 20ms/step - loss: 0.0513 - mse: 0.0047 - val_loss: 0.0808 - val_mse: 0.0103\n",
      "Epoch 118/200\n",
      "764/764 [==============================] - 15s 20ms/step - loss: 0.0518 - mse: 0.0047 - val_loss: 0.0849 - val_mse: 0.0126\n",
      "Epoch 119/200\n",
      "764/764 [==============================] - 15s 19ms/step - loss: 0.0505 - mse: 0.0046 - val_loss: 0.0913 - val_mse: 0.0134\n",
      "Epoch 120/200\n",
      "764/764 [==============================] - 15s 20ms/step - loss: 0.0500 - mse: 0.0044 - val_loss: 0.0873 - val_mse: 0.0132\n",
      "Epoch 121/200\n",
      "764/764 [==============================] - 15s 20ms/step - loss: 0.0504 - mse: 0.0045 - val_loss: 0.0720 - val_mse: 0.0088\n",
      "Epoch 122/200\n",
      "764/764 [==============================] - 16s 21ms/step - loss: 0.0505 - mse: 0.0045 - val_loss: 0.0846 - val_mse: 0.0117\n"
     ]
    }
   ],
   "source": [
    "X_train, X_val, X_test, y_train, y_val, y_test, scaler,yscaler= create_sequences_3(df, window_size=window_size_4, test_size=0.2)\n",
    "\n",
    "model_4 = Sequential([\n",
    "    GRU(64, return_sequences=True, input_shape=(window_size_4, num_features)),\n",
    "    Dropout(0.3),\n",
    "    GRU(32, return_sequences=True),\n",
    "    Dropout(0.2),\n",
    "    GRU(16, return_sequences=False),\n",
    "    Dropout(0.1),\n",
    "    Dense(32, activation='relu'),\n",
    "    Dense(1)  # Output layer for regression\n",
    "])\n",
    "model_4.compile(optimizer=Adam(learning_rate=0.001),loss='mae',metrics=['mse'])\n",
    "# Train the model\n",
    "history_4 = model_4.fit(X_train,y_train,\n",
    "                    validation_data=(X_val,y_val),\n",
    "                    epochs=200, batch_size=16,\n",
    "                    callbacks=tf.keras.callbacks.EarlyStopping(patience=30,monitor='val_loss',restore_best_weights=True))\n",
    "model_4.save('model_4.keras')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\Users\\anush\\.conda\\envs\\mlEnv_2\\lib\\site-packages\\sklearn\\preprocessing\\_data.py:1020: RuntimeWarning: invalid value encountered in sqrt\n",
      "  np.sqrt(self.var_), copy=False, constant_mask=constant_mask\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/200\n",
      "721/721 [==============================] - 21s 23ms/step - loss: 0.6266 - mse: 0.6092 - val_loss: 0.6213 - val_mse: 0.6139\n",
      "Epoch 2/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.6010 - mse: 0.5697 - val_loss: 0.5572 - val_mse: 0.5081\n",
      "Epoch 3/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.5400 - mse: 0.4818 - val_loss: 0.4788 - val_mse: 0.3799\n",
      "Epoch 4/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.4469 - mse: 0.3552 - val_loss: 0.3753 - val_mse: 0.2615\n",
      "Epoch 5/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.3610 - mse: 0.2429 - val_loss: 0.3138 - val_mse: 0.1854\n",
      "Epoch 6/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.3000 - mse: 0.1715 - val_loss: 0.2492 - val_mse: 0.1247\n",
      "Epoch 7/200\n",
      "721/721 [==============================] - 15s 20ms/step - loss: 0.2515 - mse: 0.1194 - val_loss: 0.2095 - val_mse: 0.0870\n",
      "Epoch 8/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.2149 - mse: 0.0861 - val_loss: 0.1962 - val_mse: 0.0691\n",
      "Epoch 9/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.1838 - mse: 0.0631 - val_loss: 0.1484 - val_mse: 0.0419\n",
      "Epoch 10/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.1686 - mse: 0.0515 - val_loss: 0.1611 - val_mse: 0.0450\n",
      "Epoch 11/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.1540 - mse: 0.0430 - val_loss: 0.1192 - val_mse: 0.0264\n",
      "Epoch 12/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.1436 - mse: 0.0371 - val_loss: 0.1380 - val_mse: 0.0330\n",
      "Epoch 13/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.1334 - mse: 0.0319 - val_loss: 0.1080 - val_mse: 0.0217\n",
      "Epoch 14/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.1276 - mse: 0.0298 - val_loss: 0.1059 - val_mse: 0.0210\n",
      "Epoch 15/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.1208 - mse: 0.0256 - val_loss: 0.0906 - val_mse: 0.0155\n",
      "Epoch 16/200\n",
      "721/721 [==============================] - 14s 20ms/step - loss: 0.1164 - mse: 0.0247 - val_loss: 0.0944 - val_mse: 0.0164\n",
      "Epoch 17/200\n",
      "721/721 [==============================] - 14s 20ms/step - loss: 0.1114 - mse: 0.0221 - val_loss: 0.1108 - val_mse: 0.0211\n",
      "Epoch 18/200\n",
      "721/721 [==============================] - 14s 20ms/step - loss: 0.1049 - mse: 0.0196 - val_loss: 0.1146 - val_mse: 0.0237\n",
      "Epoch 19/200\n",
      "721/721 [==============================] - 15s 20ms/step - loss: 0.1058 - mse: 0.0200 - val_loss: 0.0995 - val_mse: 0.0182\n",
      "Epoch 20/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0992 - mse: 0.0177 - val_loss: 0.0802 - val_mse: 0.0116\n",
      "Epoch 21/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.1005 - mse: 0.0181 - val_loss: 0.0715 - val_mse: 0.0090\n",
      "Epoch 22/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.0949 - mse: 0.0161 - val_loss: 0.0773 - val_mse: 0.0108\n",
      "Epoch 23/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.0950 - mse: 0.0165 - val_loss: 0.0825 - val_mse: 0.0121\n",
      "Epoch 24/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0930 - mse: 0.0156 - val_loss: 0.1021 - val_mse: 0.0177\n",
      "Epoch 25/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0896 - mse: 0.0143 - val_loss: 0.0912 - val_mse: 0.0139\n",
      "Epoch 26/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0893 - mse: 0.0143 - val_loss: 0.0903 - val_mse: 0.0137\n",
      "Epoch 27/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.0860 - mse: 0.0135 - val_loss: 0.0933 - val_mse: 0.0139\n",
      "Epoch 28/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0867 - mse: 0.0134 - val_loss: 0.0791 - val_mse: 0.0117\n",
      "Epoch 29/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0845 - mse: 0.0127 - val_loss: 0.0682 - val_mse: 0.0082\n",
      "Epoch 30/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0810 - mse: 0.0118 - val_loss: 0.0643 - val_mse: 0.0071\n",
      "Epoch 31/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.0795 - mse: 0.0114 - val_loss: 0.0847 - val_mse: 0.0114\n",
      "Epoch 32/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0801 - mse: 0.0118 - val_loss: 0.0750 - val_mse: 0.0094\n",
      "Epoch 33/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0761 - mse: 0.0103 - val_loss: 0.0619 - val_mse: 0.0066\n",
      "Epoch 34/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0762 - mse: 0.0103 - val_loss: 0.0782 - val_mse: 0.0102\n",
      "Epoch 35/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0744 - mse: 0.0101 - val_loss: 0.0735 - val_mse: 0.0097\n",
      "Epoch 36/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0748 - mse: 0.0101 - val_loss: 0.0664 - val_mse: 0.0074\n",
      "Epoch 37/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0716 - mse: 0.0094 - val_loss: 0.0742 - val_mse: 0.0091\n",
      "Epoch 38/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0703 - mse: 0.0090 - val_loss: 0.0605 - val_mse: 0.0060\n",
      "Epoch 39/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0708 - mse: 0.0090 - val_loss: 0.0658 - val_mse: 0.0074\n",
      "Epoch 40/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0700 - mse: 0.0088 - val_loss: 0.1095 - val_mse: 0.0213\n",
      "Epoch 41/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0697 - mse: 0.0089 - val_loss: 0.0864 - val_mse: 0.0116\n",
      "Epoch 42/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.0681 - mse: 0.0085 - val_loss: 0.0802 - val_mse: 0.0102\n",
      "Epoch 43/200\n",
      "721/721 [==============================] - 17s 24ms/step - loss: 0.0668 - mse: 0.0081 - val_loss: 0.0962 - val_mse: 0.0154\n",
      "Epoch 44/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.0660 - mse: 0.0078 - val_loss: 0.0593 - val_mse: 0.0064\n",
      "Epoch 45/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.0657 - mse: 0.0079 - val_loss: 0.0710 - val_mse: 0.0082\n",
      "Epoch 46/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0643 - mse: 0.0075 - val_loss: 0.0878 - val_mse: 0.0118\n",
      "Epoch 47/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0637 - mse: 0.0075 - val_loss: 0.0835 - val_mse: 0.0115\n",
      "Epoch 48/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0626 - mse: 0.0072 - val_loss: 0.0728 - val_mse: 0.0088\n",
      "Epoch 49/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0615 - mse: 0.0069 - val_loss: 0.0815 - val_mse: 0.0108\n",
      "Epoch 50/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0624 - mse: 0.0071 - val_loss: 0.0797 - val_mse: 0.0098\n",
      "Epoch 51/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0609 - mse: 0.0070 - val_loss: 0.0683 - val_mse: 0.0079\n",
      "Epoch 52/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0598 - mse: 0.0067 - val_loss: 0.0624 - val_mse: 0.0065\n",
      "Epoch 53/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0598 - mse: 0.0065 - val_loss: 0.0596 - val_mse: 0.0060\n",
      "Epoch 54/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0597 - mse: 0.0065 - val_loss: 0.0859 - val_mse: 0.0124\n",
      "Epoch 55/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0591 - mse: 0.0064 - val_loss: 0.0674 - val_mse: 0.0079\n",
      "Epoch 56/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.0588 - mse: 0.0064 - val_loss: 0.0659 - val_mse: 0.0068\n",
      "Epoch 57/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0571 - mse: 0.0061 - val_loss: 0.0794 - val_mse: 0.0104\n",
      "Epoch 58/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0572 - mse: 0.0060 - val_loss: 0.0575 - val_mse: 0.0056\n",
      "Epoch 59/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0575 - mse: 0.0062 - val_loss: 0.0773 - val_mse: 0.0105\n",
      "Epoch 60/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.0555 - mse: 0.0057 - val_loss: 0.0702 - val_mse: 0.0083\n",
      "Epoch 61/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.0555 - mse: 0.0058 - val_loss: 0.0616 - val_mse: 0.0073\n",
      "Epoch 62/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.0543 - mse: 0.0054 - val_loss: 0.0682 - val_mse: 0.0078\n",
      "Epoch 63/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.0553 - mse: 0.0058 - val_loss: 0.0766 - val_mse: 0.0102\n",
      "Epoch 64/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.0550 - mse: 0.0057 - val_loss: 0.0808 - val_mse: 0.0103\n",
      "Epoch 65/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.0526 - mse: 0.0052 - val_loss: 0.0772 - val_mse: 0.0093\n",
      "Epoch 66/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.0536 - mse: 0.0055 - val_loss: 0.0548 - val_mse: 0.0047\n",
      "Epoch 67/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.0526 - mse: 0.0051 - val_loss: 0.0913 - val_mse: 0.0126\n",
      "Epoch 68/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.0525 - mse: 0.0052 - val_loss: 0.0829 - val_mse: 0.0109\n",
      "Epoch 69/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.0586 - mse: 0.0071 - val_loss: 0.0761 - val_mse: 0.0099\n",
      "Epoch 70/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.0510 - mse: 0.0049 - val_loss: 0.0812 - val_mse: 0.0103\n",
      "Epoch 71/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.0505 - mse: 0.0048 - val_loss: 0.0648 - val_mse: 0.0071\n",
      "Epoch 72/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.0495 - mse: 0.0046 - val_loss: 0.0725 - val_mse: 0.0083\n",
      "Epoch 73/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0501 - mse: 0.0050 - val_loss: 0.0801 - val_mse: 0.0094\n",
      "Epoch 74/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0498 - mse: 0.0046 - val_loss: 0.0766 - val_mse: 0.0093\n",
      "Epoch 75/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.0516 - mse: 0.0053 - val_loss: 0.0822 - val_mse: 0.0115\n",
      "Epoch 76/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.0477 - mse: 0.0042 - val_loss: 0.0757 - val_mse: 0.0091\n",
      "Epoch 77/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0480 - mse: 0.0044 - val_loss: 0.0831 - val_mse: 0.0115\n",
      "Epoch 78/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0483 - mse: 0.0045 - val_loss: 0.0683 - val_mse: 0.0073\n",
      "Epoch 79/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0482 - mse: 0.0045 - val_loss: 0.0771 - val_mse: 0.0095\n",
      "Epoch 80/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.0482 - mse: 0.0043 - val_loss: 0.0713 - val_mse: 0.0081\n",
      "Epoch 81/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0503 - mse: 0.0052 - val_loss: 0.0699 - val_mse: 0.0073\n",
      "Epoch 82/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0473 - mse: 0.0042 - val_loss: 0.0667 - val_mse: 0.0073\n",
      "Epoch 83/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0466 - mse: 0.0041 - val_loss: 0.0539 - val_mse: 0.0048\n",
      "Epoch 84/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0466 - mse: 0.0040 - val_loss: 0.0541 - val_mse: 0.0051\n",
      "Epoch 85/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0474 - mse: 0.0043 - val_loss: 0.0766 - val_mse: 0.0096\n",
      "Epoch 86/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0461 - mse: 0.0041 - val_loss: 0.0764 - val_mse: 0.0097\n",
      "Epoch 87/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0464 - mse: 0.0042 - val_loss: 0.0766 - val_mse: 0.0098\n",
      "Epoch 88/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.0456 - mse: 0.0039 - val_loss: 0.0698 - val_mse: 0.0081\n",
      "Epoch 89/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0451 - mse: 0.0040 - val_loss: 0.0792 - val_mse: 0.0107\n",
      "Epoch 90/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.0457 - mse: 0.0040 - val_loss: 0.0745 - val_mse: 0.0094\n",
      "Epoch 91/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.0446 - mse: 0.0039 - val_loss: 0.0653 - val_mse: 0.0070\n",
      "Epoch 92/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.0447 - mse: 0.0039 - val_loss: 0.0743 - val_mse: 0.0088\n",
      "Epoch 93/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.0449 - mse: 0.0039 - val_loss: 0.0734 - val_mse: 0.0093\n",
      "Epoch 94/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0440 - mse: 0.0037 - val_loss: 0.0805 - val_mse: 0.0105\n",
      "Epoch 95/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.0442 - mse: 0.0039 - val_loss: 0.0646 - val_mse: 0.0076\n",
      "Epoch 96/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.0430 - mse: 0.0036 - val_loss: 0.0716 - val_mse: 0.0085\n",
      "Epoch 97/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.0440 - mse: 0.0039 - val_loss: 0.0843 - val_mse: 0.0119\n",
      "Epoch 98/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.0440 - mse: 0.0038 - val_loss: 0.0840 - val_mse: 0.0117\n",
      "Epoch 99/200\n",
      "721/721 [==============================] - 16s 22ms/step - loss: 0.0425 - mse: 0.0036 - val_loss: 0.0652 - val_mse: 0.0068\n",
      "Epoch 100/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0428 - mse: 0.0034 - val_loss: 0.0646 - val_mse: 0.0078\n",
      "Epoch 101/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0422 - mse: 0.0034 - val_loss: 0.0760 - val_mse: 0.0108\n",
      "Epoch 102/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0410 - mse: 0.0032 - val_loss: 0.0652 - val_mse: 0.0078\n",
      "Epoch 103/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0420 - mse: 0.0035 - val_loss: 0.0764 - val_mse: 0.0092\n",
      "Epoch 104/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0422 - mse: 0.0034 - val_loss: 0.0946 - val_mse: 0.0125\n",
      "Epoch 105/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0414 - mse: 0.0034 - val_loss: 0.0594 - val_mse: 0.0065\n",
      "Epoch 106/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0411 - mse: 0.0033 - val_loss: 0.0723 - val_mse: 0.0086\n",
      "Epoch 107/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0409 - mse: 0.0032 - val_loss: 0.0806 - val_mse: 0.0107\n",
      "Epoch 108/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0410 - mse: 0.0032 - val_loss: 0.0641 - val_mse: 0.0079\n",
      "Epoch 109/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0398 - mse: 0.0031 - val_loss: 0.0821 - val_mse: 0.0115\n",
      "Epoch 110/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0399 - mse: 0.0031 - val_loss: 0.0722 - val_mse: 0.0085\n",
      "Epoch 111/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0398 - mse: 0.0032 - val_loss: 0.0775 - val_mse: 0.0097\n",
      "Epoch 112/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0401 - mse: 0.0031 - val_loss: 0.0779 - val_mse: 0.0099\n",
      "Epoch 113/200\n",
      "721/721 [==============================] - 15s 21ms/step - loss: 0.0403 - mse: 0.0034 - val_loss: 0.0719 - val_mse: 0.0077\n"
     ]
    }
   ],
   "source": [
    "X_train, X_val, X_test, y_train, y_val, y_test, scaler,yscaler= create_sequences_3(df, window_size=window_size_5, test_size=0.2)\n",
    "\n",
    "model_5 = Sequential([\n",
    "    GRU(64, return_sequences=True, input_shape=(window_size_5, num_features)),\n",
    "    Dropout(0.3),\n",
    "    GRU(32, return_sequences=True),\n",
    "    Dropout(0.2),\n",
    "    GRU(16, return_sequences=False),\n",
    "    Dropout(0.1),\n",
    "    Dense(32, activation='relu'),\n",
    "    Dense(1)  # Output layer for regression\n",
    "])\n",
    "model_5.compile(optimizer=Adam(learning_rate=0.001),loss='mae',metrics=['mse'])\n",
    "# Train the model\n",
    "history_5 = model_5.fit(X_train,y_train,\n",
    "                    validation_data=(X_val,y_val),\n",
    "                    epochs=200, batch_size=16,\n",
    "                    callbacks=tf.keras.callbacks.EarlyStopping(patience=30,monitor='val_loss',restore_best_weights=True))\n",
    "model_5.save('model_5.keras')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/200\n",
      "678/678 [==============================] - 21s 25ms/step - loss: 0.6333 - mse: 0.6232 - val_loss: 0.6272 - val_mse: 0.6080\n",
      "Epoch 2/200\n",
      "678/678 [==============================] - 16s 23ms/step - loss: 0.5892 - mse: 0.5549 - val_loss: 0.5715 - val_mse: 0.5303\n",
      "Epoch 3/200\n",
      "678/678 [==============================] - 16s 23ms/step - loss: 0.4964 - mse: 0.4206 - val_loss: 0.4203 - val_mse: 0.3174\n",
      "Epoch 4/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.3937 - mse: 0.2798 - val_loss: 0.3144 - val_mse: 0.1918\n",
      "Epoch 5/200\n",
      "678/678 [==============================] - 16s 23ms/step - loss: 0.3078 - mse: 0.1775 - val_loss: 0.2578 - val_mse: 0.1280\n",
      "Epoch 6/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.2522 - mse: 0.1193 - val_loss: 0.1922 - val_mse: 0.0704\n",
      "Epoch 7/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.2031 - mse: 0.0778 - val_loss: 0.1694 - val_mse: 0.0564\n",
      "Epoch 8/200\n",
      "678/678 [==============================] - 17s 25ms/step - loss: 0.1773 - mse: 0.0590 - val_loss: 0.1407 - val_mse: 0.0387\n",
      "Epoch 9/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.1593 - mse: 0.0468 - val_loss: 0.1287 - val_mse: 0.0337\n",
      "Epoch 10/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.1441 - mse: 0.0382 - val_loss: 0.1112 - val_mse: 0.0253\n",
      "Epoch 11/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.1342 - mse: 0.0334 - val_loss: 0.1171 - val_mse: 0.0262\n",
      "Epoch 12/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.1259 - mse: 0.0293 - val_loss: 0.1284 - val_mse: 0.0313\n",
      "Epoch 13/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.1174 - mse: 0.0252 - val_loss: 0.1046 - val_mse: 0.0211\n",
      "Epoch 14/200\n",
      "678/678 [==============================] - 17s 24ms/step - loss: 0.1129 - mse: 0.0233 - val_loss: 0.1087 - val_mse: 0.0228\n",
      "Epoch 15/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.1105 - mse: 0.0221 - val_loss: 0.0871 - val_mse: 0.0144\n",
      "Epoch 16/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.1042 - mse: 0.0200 - val_loss: 0.0935 - val_mse: 0.0164\n",
      "Epoch 17/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.0993 - mse: 0.0178 - val_loss: 0.0965 - val_mse: 0.0182\n",
      "Epoch 18/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.0986 - mse: 0.0176 - val_loss: 0.0855 - val_mse: 0.0131\n",
      "Epoch 19/200\n",
      "678/678 [==============================] - 16s 23ms/step - loss: 0.0977 - mse: 0.0171 - val_loss: 0.0799 - val_mse: 0.0108\n",
      "Epoch 20/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.0920 - mse: 0.0153 - val_loss: 0.1074 - val_mse: 0.0193\n",
      "Epoch 21/200\n",
      "678/678 [==============================] - 16s 23ms/step - loss: 0.0909 - mse: 0.0152 - val_loss: 0.1004 - val_mse: 0.0172\n",
      "Epoch 22/200\n",
      "678/678 [==============================] - 16s 23ms/step - loss: 0.0879 - mse: 0.0139 - val_loss: 0.0925 - val_mse: 0.0143\n",
      "Epoch 23/200\n",
      "678/678 [==============================] - 16s 23ms/step - loss: 0.0866 - mse: 0.0136 - val_loss: 0.0835 - val_mse: 0.0128\n",
      "Epoch 24/200\n",
      "678/678 [==============================] - 16s 23ms/step - loss: 0.0848 - mse: 0.0131 - val_loss: 0.0822 - val_mse: 0.0130\n",
      "Epoch 25/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.0808 - mse: 0.0120 - val_loss: 0.0711 - val_mse: 0.0091\n",
      "Epoch 26/200\n",
      "678/678 [==============================] - 15s 23ms/step - loss: 0.0792 - mse: 0.0113 - val_loss: 0.0671 - val_mse: 0.0077\n",
      "Epoch 27/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.0786 - mse: 0.0112 - val_loss: 0.0716 - val_mse: 0.0081\n",
      "Epoch 28/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.0779 - mse: 0.0108 - val_loss: 0.0841 - val_mse: 0.0114\n",
      "Epoch 29/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.0751 - mse: 0.0101 - val_loss: 0.1083 - val_mse: 0.0175\n",
      "Epoch 30/200\n",
      "678/678 [==============================] - 16s 23ms/step - loss: 0.0754 - mse: 0.0101 - val_loss: 0.0779 - val_mse: 0.0107\n",
      "Epoch 31/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.0735 - mse: 0.0097 - val_loss: 0.0865 - val_mse: 0.0126\n",
      "Epoch 32/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.0712 - mse: 0.0091 - val_loss: 0.0817 - val_mse: 0.0112\n",
      "Epoch 33/200\n",
      "678/678 [==============================] - 17s 25ms/step - loss: 0.0703 - mse: 0.0089 - val_loss: 0.1040 - val_mse: 0.0172\n",
      "Epoch 34/200\n",
      "678/678 [==============================] - 17s 25ms/step - loss: 0.0680 - mse: 0.0085 - val_loss: 0.0538 - val_mse: 0.0054\n",
      "Epoch 35/200\n",
      "678/678 [==============================] - 16s 23ms/step - loss: 0.0686 - mse: 0.0083 - val_loss: 0.0676 - val_mse: 0.0077\n",
      "Epoch 36/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.0676 - mse: 0.0083 - val_loss: 0.0987 - val_mse: 0.0160\n",
      "Epoch 37/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.0636 - mse: 0.0073 - val_loss: 0.0829 - val_mse: 0.0109\n",
      "Epoch 38/200\n",
      "678/678 [==============================] - 15s 22ms/step - loss: 0.0630 - mse: 0.0072 - val_loss: 0.0790 - val_mse: 0.0100\n",
      "Epoch 39/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.0630 - mse: 0.0071 - val_loss: 0.0932 - val_mse: 0.0142\n",
      "Epoch 40/200\n",
      "678/678 [==============================] - 15s 23ms/step - loss: 0.0625 - mse: 0.0072 - val_loss: 0.0675 - val_mse: 0.0075\n",
      "Epoch 41/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.0614 - mse: 0.0068 - val_loss: 0.0677 - val_mse: 0.0079\n",
      "Epoch 42/200\n",
      "678/678 [==============================] - 16s 23ms/step - loss: 0.0607 - mse: 0.0065 - val_loss: 0.0939 - val_mse: 0.0149\n",
      "Epoch 43/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.0608 - mse: 0.0067 - val_loss: 0.0615 - val_mse: 0.0069\n",
      "Epoch 44/200\n",
      "678/678 [==============================] - 15s 22ms/step - loss: 0.0589 - mse: 0.0063 - val_loss: 0.0990 - val_mse: 0.0161\n",
      "Epoch 45/200\n",
      "678/678 [==============================] - 16s 23ms/step - loss: 0.0584 - mse: 0.0062 - val_loss: 0.0794 - val_mse: 0.0102\n",
      "Epoch 46/200\n",
      "678/678 [==============================] - 16s 23ms/step - loss: 0.0583 - mse: 0.0062 - val_loss: 0.0564 - val_mse: 0.0056\n",
      "Epoch 47/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.0572 - mse: 0.0060 - val_loss: 0.1000 - val_mse: 0.0169\n",
      "Epoch 48/200\n",
      "678/678 [==============================] - 15s 23ms/step - loss: 0.0561 - mse: 0.0058 - val_loss: 0.0699 - val_mse: 0.0078\n",
      "Epoch 49/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.0553 - mse: 0.0055 - val_loss: 0.0726 - val_mse: 0.0086\n",
      "Epoch 50/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.0551 - mse: 0.0057 - val_loss: 0.0683 - val_mse: 0.0078\n",
      "Epoch 51/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.0549 - mse: 0.0055 - val_loss: 0.0694 - val_mse: 0.0086\n",
      "Epoch 52/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.0543 - mse: 0.0055 - val_loss: 0.0784 - val_mse: 0.0087\n",
      "Epoch 53/200\n",
      "678/678 [==============================] - 16s 23ms/step - loss: 0.0530 - mse: 0.0051 - val_loss: 0.0797 - val_mse: 0.0097\n",
      "Epoch 54/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.0534 - mse: 0.0053 - val_loss: 0.0929 - val_mse: 0.0126\n",
      "Epoch 55/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.0517 - mse: 0.0049 - val_loss: 0.0661 - val_mse: 0.0073\n",
      "Epoch 56/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.0517 - mse: 0.0049 - val_loss: 0.0817 - val_mse: 0.0116\n",
      "Epoch 57/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.0507 - mse: 0.0049 - val_loss: 0.0596 - val_mse: 0.0057\n",
      "Epoch 58/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.0502 - mse: 0.0047 - val_loss: 0.0702 - val_mse: 0.0080\n",
      "Epoch 59/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.0493 - mse: 0.0046 - val_loss: 0.0685 - val_mse: 0.0081\n",
      "Epoch 60/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.0497 - mse: 0.0045 - val_loss: 0.0924 - val_mse: 0.0141\n",
      "Epoch 61/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.0499 - mse: 0.0047 - val_loss: 0.0655 - val_mse: 0.0068\n",
      "Epoch 62/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.0487 - mse: 0.0044 - val_loss: 0.0869 - val_mse: 0.0112\n",
      "Epoch 63/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.0483 - mse: 0.0044 - val_loss: 0.0801 - val_mse: 0.0100\n",
      "Epoch 64/200\n",
      "678/678 [==============================] - 16s 24ms/step - loss: 0.0488 - mse: 0.0044 - val_loss: 0.0959 - val_mse: 0.0135\n"
     ]
    }
   ],
   "source": [
    "X_train, X_val, X_test, y_train, y_val, y_test, scaler,yscaler= create_sequences_3(df, window_size=window_size_6, test_size=0.2)\n",
    "\n",
    "model_6 = Sequential([\n",
    "    GRU(64, return_sequences=True, input_shape=(window_size_6, num_features)),\n",
    "    Dropout(0.3),\n",
    "    GRU(32, return_sequences=True),\n",
    "    Dropout(0.2),\n",
    "    GRU(16, return_sequences=False),\n",
    "    Dropout(0.1),\n",
    "    Dense(32, activation='relu'),\n",
    "    Dense(1)  # Output layer for regression\n",
    "])\n",
    "model_6.compile(optimizer=Adam(learning_rate=0.001),loss='mae',metrics=['mse'])\n",
    "# Train the model\n",
    "history_6 = model_6.fit(X_train,y_train,\n",
    "                   validation_data=(X_val,y_val),\n",
    "                    epochs=200, batch_size=16,\n",
    "                    callbacks=tf.keras.callbacks.EarlyStopping(patience=30,monitor='val_loss',restore_best_weights=True))\n",
    "model_6.save('model_6.keras')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "window_len 5 model Model Evaluation - RMSE: 51.62, MAE: 38.46\n"
     ]
    }
   ],
   "source": [
    "X_train, X_val, X_test, y_train, y_val, y_test, scaler,yscaler= create_sequences_3(df, window_size=window_size_1, test_size=0.2)\n",
    "metrics_results.append(evaluate_model(model_1, X_test,y_test,f'window_len {window_size_1} model',yscaler))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "window_len 10 model Model Evaluation - RMSE: 23.09, MAE: 15.74\n"
     ]
    }
   ],
   "source": [
    "X_train, X_val, X_test, y_train, y_val, y_test, scaler,yscaler= create_sequences_3(df, window_size=window_size_2, test_size=0.2)\n",
    "metrics_results.append(evaluate_model(model_2, X_test,y_test,f'window_len {window_size_2} model',yscaler))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "window_len 20 model Model Evaluation - RMSE: 8.00, MAE: 5.83\n"
     ]
    }
   ],
   "source": [
    "X_train, X_val, X_test, y_train, y_val, y_test, scaler,yscaler= create_sequences_3(df, window_size=window_size_3, test_size=0.2)\n",
    "metrics_results.append(evaluate_model(model_3, X_test,y_test,f'window_len {window_size_3} model',yscaler))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "window_len 30 model Model Evaluation - RMSE: 6.43, MAE: 4.95\n"
     ]
    }
   ],
   "source": [
    "X_train, X_val, X_test, y_train, y_val, y_test, scaler,yscaler= create_sequences_3(df, window_size=window_size_4, test_size=0.2)\n",
    "metrics_results.append(evaluate_model(model_4, X_test,y_test,f'window_len {window_size_4} model',yscaler))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\Users\\anush\\.conda\\envs\\mlEnv_2\\lib\\site-packages\\sklearn\\preprocessing\\_data.py:1020: RuntimeWarning: invalid value encountered in sqrt\n",
      "  np.sqrt(self.var_), copy=False, constant_mask=constant_mask\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "window_len 40 model Model Evaluation - RMSE: 5.44, MAE: 4.12\n"
     ]
    }
   ],
   "source": [
    "X_train, X_val, X_test, y_train, y_val, y_test, scaler,yscaler= create_sequences_3(df, window_size=window_size_5, test_size=0.2)\n",
    "metrics_results.append(evaluate_model(model_5, X_test,y_test,f'window_len {window_size_5} model',yscaler))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "window_len 50 model Model Evaluation - RMSE: 5.27, MAE: 3.95\n"
     ]
    }
   ],
   "source": [
    "X_train, X_val, X_test, y_train, y_val, y_test, scaler,yscaler= create_sequences_3(df, window_size=window_size_6, test_size=0.2)\n",
    "metrics_results.append(evaluate_model(model_6, X_test,y_test,f'window_len {window_size_6} model',yscaler))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "window_lengths = [5,10,20,30,40,50]\n",
    "\n",
    "def plot_metrics_vs_window(results, window_lengths):\n",
    "    \"\"\"\n",
    "    Plots MAE and RMSE vs. window lengths.\n",
    "    \n",
    "    Parameters:\n",
    "        results (list of tuples): Each tuple contains (rmse, mae) for a model.\n",
    "        window_lengths (list): List of window lengths corresponding to each result.\n",
    "    \"\"\"\n",
    "    # Extract RMSE and MAE from results\n",
    "    rmse_values = [result[1] for result in results]\n",
    "    mae_values = [result[2] for result in results]\n",
    "    \n",
    "    # Plotting\n",
    "    plt.figure(figsize=(10, 6))\n",
    "    plt.plot(window_lengths, rmse_values, label=\"RMSE\", marker='o', color='b')\n",
    "    plt.plot(window_lengths, mae_values, label=\"MAE\", marker='o', color='r')\n",
    "    \n",
    "    plt.title(\"Model Performance vs. Window Length\")\n",
    "    plt.xlabel(\"Window Length\")\n",
    "    plt.ylabel(\"Error\")\n",
    "    plt.legend()\n",
    "    plt.grid(True)\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_metrics_vs_window(metrics_results, window_lengths)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "mlEnv_2",
   "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": "3.9.20"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}