LSTM-Power-Demand-Forecasting.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "6b3adbae-64c8-47f9-9177-e062d693a370",
   "metadata": {},
   "source": [
    "# POWER DEMAND FORECASTING USING LSTM\n",
    "## Code By: Abtahi Ishmam, Koushik Roy\n",
    "### Paper Link: https://doi.org/10.1109/ACMI53878.2021.9528277"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4b8166ab-e799-4379-93e9-b5a9541de5ba",
   "metadata": {},
   "source": [
    "### Step 1: Importing the necessary libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "bd186c37-ed49-4d5e-adbe-6a218ec1cc24",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "01a4df66-d4fd-42c2-ad32-d8e4141b9679",
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers import Dense\n",
    "from keras.layers import LSTM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "259d9113-7b62-444a-aa53-c8350baffa48",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import MinMaxScaler\n",
    "from sklearn.model_selection import train_test_split"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "367b338f-0b1e-47ce-b99c-705d99fdba2c",
   "metadata": {},
   "source": [
    "### Step 2: Dataset exploration and preprocessing"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a7d7a223-523e-4142-8e01-065885de3a92",
   "metadata": {},
   "source": [
    "#### 2a. Exploration"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "42d29dc9-1d7f-4050-ba9e-0017a8bfc379",
   "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>time</th>\n",
       "      <th>generation biomass</th>\n",
       "      <th>generation fossil brown coal/lignite</th>\n",
       "      <th>generation fossil coal-derived gas</th>\n",
       "      <th>generation fossil gas</th>\n",
       "      <th>generation fossil hard coal</th>\n",
       "      <th>generation fossil oil</th>\n",
       "      <th>generation fossil oil shale</th>\n",
       "      <th>generation fossil peat</th>\n",
       "      <th>generation geothermal</th>\n",
       "      <th>...</th>\n",
       "      <th>generation waste</th>\n",
       "      <th>generation wind offshore</th>\n",
       "      <th>generation wind onshore</th>\n",
       "      <th>forecast solar day ahead</th>\n",
       "      <th>forecast wind offshore eday ahead</th>\n",
       "      <th>forecast wind onshore day ahead</th>\n",
       "      <th>total load forecast</th>\n",
       "      <th>total load actual</th>\n",
       "      <th>price day ahead</th>\n",
       "      <th>price actual</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2015-01-01 00:00:00+01:00</td>\n",
       "      <td>447.0</td>\n",
       "      <td>329.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4844.0</td>\n",
       "      <td>4821.0</td>\n",
       "      <td>162.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>196.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>6378.0</td>\n",
       "      <td>17.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>6436.0</td>\n",
       "      <td>26118.0</td>\n",
       "      <td>25385.0</td>\n",
       "      <td>50.10</td>\n",
       "      <td>65.41</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2015-01-01 01:00:00+01:00</td>\n",
       "      <td>449.0</td>\n",
       "      <td>328.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5196.0</td>\n",
       "      <td>4755.0</td>\n",
       "      <td>158.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>195.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5890.0</td>\n",
       "      <td>16.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5856.0</td>\n",
       "      <td>24934.0</td>\n",
       "      <td>24382.0</td>\n",
       "      <td>48.10</td>\n",
       "      <td>64.92</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2015-01-01 02:00:00+01:00</td>\n",
       "      <td>448.0</td>\n",
       "      <td>323.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4857.0</td>\n",
       "      <td>4581.0</td>\n",
       "      <td>157.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>196.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5461.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5454.0</td>\n",
       "      <td>23515.0</td>\n",
       "      <td>22734.0</td>\n",
       "      <td>47.33</td>\n",
       "      <td>64.48</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2015-01-01 03:00:00+01:00</td>\n",
       "      <td>438.0</td>\n",
       "      <td>254.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4314.0</td>\n",
       "      <td>4131.0</td>\n",
       "      <td>160.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>191.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5238.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5151.0</td>\n",
       "      <td>22642.0</td>\n",
       "      <td>21286.0</td>\n",
       "      <td>42.27</td>\n",
       "      <td>59.32</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2015-01-01 04:00:00+01:00</td>\n",
       "      <td>428.0</td>\n",
       "      <td>187.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4130.0</td>\n",
       "      <td>3840.0</td>\n",
       "      <td>156.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>189.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4935.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4861.0</td>\n",
       "      <td>21785.0</td>\n",
       "      <td>20264.0</td>\n",
       "      <td>38.41</td>\n",
       "      <td>56.04</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 29 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                        time  generation biomass  \\\n",
       "0  2015-01-01 00:00:00+01:00               447.0   \n",
       "1  2015-01-01 01:00:00+01:00               449.0   \n",
       "2  2015-01-01 02:00:00+01:00               448.0   \n",
       "3  2015-01-01 03:00:00+01:00               438.0   \n",
       "4  2015-01-01 04:00:00+01:00               428.0   \n",
       "\n",
       "   generation fossil brown coal/lignite  generation fossil coal-derived gas  \\\n",
       "0                                 329.0                                 0.0   \n",
       "1                                 328.0                                 0.0   \n",
       "2                                 323.0                                 0.0   \n",
       "3                                 254.0                                 0.0   \n",
       "4                                 187.0                                 0.0   \n",
       "\n",
       "   generation fossil gas  generation fossil hard coal  generation fossil oil  \\\n",
       "0                 4844.0                       4821.0                  162.0   \n",
       "1                 5196.0                       4755.0                  158.0   \n",
       "2                 4857.0                       4581.0                  157.0   \n",
       "3                 4314.0                       4131.0                  160.0   \n",
       "4                 4130.0                       3840.0                  156.0   \n",
       "\n",
       "   generation fossil oil shale  generation fossil peat  generation geothermal  \\\n",
       "0                          0.0                     0.0                    0.0   \n",
       "1                          0.0                     0.0                    0.0   \n",
       "2                          0.0                     0.0                    0.0   \n",
       "3                          0.0                     0.0                    0.0   \n",
       "4                          0.0                     0.0                    0.0   \n",
       "\n",
       "   ...  generation waste  generation wind offshore  generation wind onshore  \\\n",
       "0  ...             196.0                       0.0                   6378.0   \n",
       "1  ...             195.0                       0.0                   5890.0   \n",
       "2  ...             196.0                       0.0                   5461.0   \n",
       "3  ...             191.0                       0.0                   5238.0   \n",
       "4  ...             189.0                       0.0                   4935.0   \n",
       "\n",
       "   forecast solar day ahead  forecast wind offshore eday ahead  \\\n",
       "0                      17.0                                NaN   \n",
       "1                      16.0                                NaN   \n",
       "2                       8.0                                NaN   \n",
       "3                       2.0                                NaN   \n",
       "4                       9.0                                NaN   \n",
       "\n",
       "   forecast wind onshore day ahead  total load forecast  total load actual  \\\n",
       "0                           6436.0              26118.0            25385.0   \n",
       "1                           5856.0              24934.0            24382.0   \n",
       "2                           5454.0              23515.0            22734.0   \n",
       "3                           5151.0              22642.0            21286.0   \n",
       "4                           4861.0              21785.0            20264.0   \n",
       "\n",
       "   price day ahead  price actual  \n",
       "0            50.10         65.41  \n",
       "1            48.10         64.92  \n",
       "2            47.33         64.48  \n",
       "3            42.27         59.32  \n",
       "4            38.41         56.04  \n",
       "\n",
       "[5 rows x 29 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.read_csv('energy_dataset.csv')\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "a963553b-c04f-4ede-b438-fed5b664ca98",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 35064 entries, 0 to 35063\n",
      "Data columns (total 29 columns):\n",
      " #   Column                                       Non-Null Count  Dtype  \n",
      "---  ------                                       --------------  -----  \n",
      " 0   time                                         35064 non-null  object \n",
      " 1   generation biomass                           35045 non-null  float64\n",
      " 2   generation fossil brown coal/lignite         35046 non-null  float64\n",
      " 3   generation fossil coal-derived gas           35046 non-null  float64\n",
      " 4   generation fossil gas                        35046 non-null  float64\n",
      " 5   generation fossil hard coal                  35046 non-null  float64\n",
      " 6   generation fossil oil                        35045 non-null  float64\n",
      " 7   generation fossil oil shale                  35046 non-null  float64\n",
      " 8   generation fossil peat                       35046 non-null  float64\n",
      " 9   generation geothermal                        35046 non-null  float64\n",
      " 10  generation hydro pumped storage aggregated   0 non-null      float64\n",
      " 11  generation hydro pumped storage consumption  35045 non-null  float64\n",
      " 12  generation hydro run-of-river and poundage   35045 non-null  float64\n",
      " 13  generation hydro water reservoir             35046 non-null  float64\n",
      " 14  generation marine                            35045 non-null  float64\n",
      " 15  generation nuclear                           35047 non-null  float64\n",
      " 16  generation other                             35046 non-null  float64\n",
      " 17  generation other renewable                   35046 non-null  float64\n",
      " 18  generation solar                             35046 non-null  float64\n",
      " 19  generation waste                             35045 non-null  float64\n",
      " 20  generation wind offshore                     35046 non-null  float64\n",
      " 21  generation wind onshore                      35046 non-null  float64\n",
      " 22  forecast solar day ahead                     35064 non-null  float64\n",
      " 23  forecast wind offshore eday ahead            0 non-null      float64\n",
      " 24  forecast wind onshore day ahead              35064 non-null  float64\n",
      " 25  total load forecast                          35064 non-null  float64\n",
      " 26  total load actual                            35028 non-null  float64\n",
      " 27  price day ahead                              35064 non-null  float64\n",
      " 28  price actual                                 35064 non-null  float64\n",
      "dtypes: float64(28), object(1)\n",
      "memory usage: 7.8+ MB\n"
     ]
    }
   ],
   "source": [
    "df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "525879e1-3a41-4e6b-ab7e-68cab35b03ec",
   "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>generation biomass</th>\n",
       "      <th>generation fossil brown coal/lignite</th>\n",
       "      <th>generation fossil coal-derived gas</th>\n",
       "      <th>generation fossil gas</th>\n",
       "      <th>generation fossil hard coal</th>\n",
       "      <th>generation fossil oil</th>\n",
       "      <th>generation fossil oil shale</th>\n",
       "      <th>generation fossil peat</th>\n",
       "      <th>generation geothermal</th>\n",
       "      <th>generation hydro pumped storage aggregated</th>\n",
       "      <th>...</th>\n",
       "      <th>generation waste</th>\n",
       "      <th>generation wind offshore</th>\n",
       "      <th>generation wind onshore</th>\n",
       "      <th>forecast solar day ahead</th>\n",
       "      <th>forecast wind offshore eday ahead</th>\n",
       "      <th>forecast wind onshore day ahead</th>\n",
       "      <th>total load forecast</th>\n",
       "      <th>total load actual</th>\n",
       "      <th>price day ahead</th>\n",
       "      <th>price actual</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>35045.000000</td>\n",
       "      <td>35046.000000</td>\n",
       "      <td>35046.0</td>\n",
       "      <td>35046.000000</td>\n",
       "      <td>35046.000000</td>\n",
       "      <td>35045.000000</td>\n",
       "      <td>35046.0</td>\n",
       "      <td>35046.0</td>\n",
       "      <td>35046.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>35045.000000</td>\n",
       "      <td>35046.0</td>\n",
       "      <td>35046.000000</td>\n",
       "      <td>35064.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>35064.000000</td>\n",
       "      <td>35064.000000</td>\n",
       "      <td>35028.000000</td>\n",
       "      <td>35064.000000</td>\n",
       "      <td>35064.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>383.513540</td>\n",
       "      <td>448.059208</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5622.737488</td>\n",
       "      <td>4256.065742</td>\n",
       "      <td>298.319789</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>269.452133</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5464.479769</td>\n",
       "      <td>1439.066735</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5471.216689</td>\n",
       "      <td>28712.129962</td>\n",
       "      <td>28696.939905</td>\n",
       "      <td>49.874341</td>\n",
       "      <td>57.884023</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>85.353943</td>\n",
       "      <td>354.568590</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2201.830478</td>\n",
       "      <td>1961.601013</td>\n",
       "      <td>52.520673</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>50.195536</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3213.691587</td>\n",
       "      <td>1677.703355</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3176.312853</td>\n",
       "      <td>4594.100854</td>\n",
       "      <td>4574.987950</td>\n",
       "      <td>14.618900</td>\n",
       "      <td>14.204083</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>237.000000</td>\n",
       "      <td>18105.000000</td>\n",
       "      <td>18041.000000</td>\n",
       "      <td>2.060000</td>\n",
       "      <td>9.330000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>333.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4126.000000</td>\n",
       "      <td>2527.000000</td>\n",
       "      <td>263.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>240.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2933.000000</td>\n",
       "      <td>69.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2979.000000</td>\n",
       "      <td>24793.750000</td>\n",
       "      <td>24807.750000</td>\n",
       "      <td>41.490000</td>\n",
       "      <td>49.347500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>367.000000</td>\n",
       "      <td>509.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4969.000000</td>\n",
       "      <td>4474.000000</td>\n",
       "      <td>300.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>279.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4849.000000</td>\n",
       "      <td>576.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4855.000000</td>\n",
       "      <td>28906.000000</td>\n",
       "      <td>28901.000000</td>\n",
       "      <td>50.520000</td>\n",
       "      <td>58.020000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>433.000000</td>\n",
       "      <td>757.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>6429.000000</td>\n",
       "      <td>5838.750000</td>\n",
       "      <td>330.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>310.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>7398.000000</td>\n",
       "      <td>2636.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>7353.000000</td>\n",
       "      <td>32263.250000</td>\n",
       "      <td>32192.000000</td>\n",
       "      <td>60.530000</td>\n",
       "      <td>68.010000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>592.000000</td>\n",
       "      <td>999.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>20034.000000</td>\n",
       "      <td>8359.000000</td>\n",
       "      <td>449.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>357.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>17436.000000</td>\n",
       "      <td>5836.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>17430.000000</td>\n",
       "      <td>41390.000000</td>\n",
       "      <td>41015.000000</td>\n",
       "      <td>101.990000</td>\n",
       "      <td>116.800000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8 rows × 28 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       generation biomass  generation fossil brown coal/lignite  \\\n",
       "count        35045.000000                          35046.000000   \n",
       "mean           383.513540                            448.059208   \n",
       "std             85.353943                            354.568590   \n",
       "min              0.000000                              0.000000   \n",
       "25%            333.000000                              0.000000   \n",
       "50%            367.000000                            509.000000   \n",
       "75%            433.000000                            757.000000   \n",
       "max            592.000000                            999.000000   \n",
       "\n",
       "       generation fossil coal-derived gas  generation fossil gas  \\\n",
       "count                             35046.0           35046.000000   \n",
       "mean                                  0.0            5622.737488   \n",
       "std                                   0.0            2201.830478   \n",
       "min                                   0.0               0.000000   \n",
       "25%                                   0.0            4126.000000   \n",
       "50%                                   0.0            4969.000000   \n",
       "75%                                   0.0            6429.000000   \n",
       "max                                   0.0           20034.000000   \n",
       "\n",
       "       generation fossil hard coal  generation fossil oil  \\\n",
       "count                 35046.000000           35045.000000   \n",
       "mean                   4256.065742             298.319789   \n",
       "std                    1961.601013              52.520673   \n",
       "min                       0.000000               0.000000   \n",
       "25%                    2527.000000             263.000000   \n",
       "50%                    4474.000000             300.000000   \n",
       "75%                    5838.750000             330.000000   \n",
       "max                    8359.000000             449.000000   \n",
       "\n",
       "       generation fossil oil shale  generation fossil peat  \\\n",
       "count                      35046.0                 35046.0   \n",
       "mean                           0.0                     0.0   \n",
       "std                            0.0                     0.0   \n",
       "min                            0.0                     0.0   \n",
       "25%                            0.0                     0.0   \n",
       "50%                            0.0                     0.0   \n",
       "75%                            0.0                     0.0   \n",
       "max                            0.0                     0.0   \n",
       "\n",
       "       generation geothermal  generation hydro pumped storage aggregated  ...  \\\n",
       "count                35046.0                                         0.0  ...   \n",
       "mean                     0.0                                         NaN  ...   \n",
       "std                      0.0                                         NaN  ...   \n",
       "min                      0.0                                         NaN  ...   \n",
       "25%                      0.0                                         NaN  ...   \n",
       "50%                      0.0                                         NaN  ...   \n",
       "75%                      0.0                                         NaN  ...   \n",
       "max                      0.0                                         NaN  ...   \n",
       "\n",
       "       generation waste  generation wind offshore  generation wind onshore  \\\n",
       "count      35045.000000                   35046.0             35046.000000   \n",
       "mean         269.452133                       0.0              5464.479769   \n",
       "std           50.195536                       0.0              3213.691587   \n",
       "min            0.000000                       0.0                 0.000000   \n",
       "25%          240.000000                       0.0              2933.000000   \n",
       "50%          279.000000                       0.0              4849.000000   \n",
       "75%          310.000000                       0.0              7398.000000   \n",
       "max          357.000000                       0.0             17436.000000   \n",
       "\n",
       "       forecast solar day ahead  forecast wind offshore eday ahead  \\\n",
       "count              35064.000000                                0.0   \n",
       "mean                1439.066735                                NaN   \n",
       "std                 1677.703355                                NaN   \n",
       "min                    0.000000                                NaN   \n",
       "25%                   69.000000                                NaN   \n",
       "50%                  576.000000                                NaN   \n",
       "75%                 2636.000000                                NaN   \n",
       "max                 5836.000000                                NaN   \n",
       "\n",
       "       forecast wind onshore day ahead  total load forecast  \\\n",
       "count                     35064.000000         35064.000000   \n",
       "mean                       5471.216689         28712.129962   \n",
       "std                        3176.312853          4594.100854   \n",
       "min                         237.000000         18105.000000   \n",
       "25%                        2979.000000         24793.750000   \n",
       "50%                        4855.000000         28906.000000   \n",
       "75%                        7353.000000         32263.250000   \n",
       "max                       17430.000000         41390.000000   \n",
       "\n",
       "       total load actual  price day ahead  price actual  \n",
       "count       35028.000000     35064.000000  35064.000000  \n",
       "mean        28696.939905        49.874341     57.884023  \n",
       "std          4574.987950        14.618900     14.204083  \n",
       "min         18041.000000         2.060000      9.330000  \n",
       "25%         24807.750000        41.490000     49.347500  \n",
       "50%         28901.000000        50.520000     58.020000  \n",
       "75%         32192.000000        60.530000     68.010000  \n",
       "max         41015.000000       101.990000    116.800000  \n",
       "\n",
       "[8 rows x 28 columns]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "bdb1f6a8-0d24-4836-9a87-25a61e1f984f",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Count the total number of null values\n",
    "total_null_values = df.isnull().sum().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "5c8a8446-13c6-428d-93e2-e456052363c4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "70529"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "total_null_values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "24241080-2241-4507-aa0e-7d09e79df328",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['time', 'generation biomass', 'generation fossil brown coal/lignite',\n",
       "       'generation fossil coal-derived gas', 'generation fossil gas',\n",
       "       'generation fossil hard coal', 'generation fossil oil',\n",
       "       'generation fossil oil shale', 'generation fossil peat',\n",
       "       'generation geothermal', 'generation hydro pumped storage aggregated',\n",
       "       'generation hydro pumped storage consumption',\n",
       "       'generation hydro run-of-river and poundage',\n",
       "       'generation hydro water reservoir', 'generation marine',\n",
       "       'generation nuclear', 'generation other', 'generation other renewable',\n",
       "       'generation solar', 'generation waste', 'generation wind offshore',\n",
       "       'generation wind onshore', 'forecast solar day ahead',\n",
       "       'forecast wind offshore eday ahead', 'forecast wind onshore day ahead',\n",
       "       'total load forecast', 'total load actual', 'price day ahead',\n",
       "       'price actual'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.columns"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "81f3a4a9-cee4-45ae-8f66-9bd20b3ae03e",
   "metadata": {},
   "source": [
    "#### --> So, there are a lot of missing data, but our main concern is the column 'load actual'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "52dfb417-d364-437b-843f-5a2d2ade23ca",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Count the total number of null values in the load_actual column\n",
    "total_null_values_in_load_actual = df['total load actual'].isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "606136bc-745b-4ab6-913e-3efcfd5dc2af",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(36, 35064)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "total_null_values_in_load_actual, len(df)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bc265913-4af5-4c62-a6ce-e1fdd29c3157",
   "metadata": {},
   "source": [
    "#### --> 36 values are missing out of total 35064 instances, We need to either fill in these values or remove the rows"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8280c591-1939-4aec-a36e-f1a1d3290468",
   "metadata": {},
   "source": [
    "#### 2b. Preprocessing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "9ab2234b-5b1b-479d-831e-f0d86cc561d3",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Missing values imputed using mean.\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "from sklearn.impute import SimpleImputer\n",
    "\n",
    "def impute_dataset(dataset, imputation_method='mean'):\n",
    "    \"\"\"\n",
    "    Impute missing values in a dataset using the specified imputation method.\n",
    "\n",
    "    Parameters:\n",
    "    - dataset (pd.DataFrame): The input dataset with missing values.\n",
    "    - imputation_method (str): The imputation method ('mean', 'median', 'mode', etc.).\n",
    "\n",
    "    Returns:\n",
    "    - pd.DataFrame: The imputed dataset.\n",
    "    \"\"\"\n",
    "\n",
    "    # Identify columns with missing values\n",
    "    columns_with_missing = dataset.columns[dataset.isnull().any()].tolist()\n",
    "\n",
    "    if not columns_with_missing:\n",
    "        print(\"No missing values found in the dataset.\")\n",
    "        return dataset\n",
    "\n",
    "    # Initialize the imputer\n",
    "    imputer = SimpleImputer(strategy=imputation_method)\n",
    "\n",
    "    # Fit and transform the imputer on the dataset\n",
    "    dataset.loc[:,columns_with_missing] = imputer.fit_transform(dataset[columns_with_missing])\n",
    "\n",
    "    print(\"Missing values imputed using {}.\".format(imputation_method))\n",
    "\n",
    "    return dataset\n",
    "\n",
    "# Impute missing values using mean imputation\n",
    "imputed_df = impute_dataset(df[['time', 'total load actual']], imputation_method='mean')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "2f7fc6d9-9e86-4b95-b28c-c9116c17c059",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Count the total number of null values\n",
    "total_null_values = imputed_df.isnull().sum().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "e422b5c0-df3d-4f77-a800-b40d642e1ddb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "total_null_values"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a88ce5ef-3de3-4d55-99a6-effaece970cc",
   "metadata": {},
   "source": [
    "#### --> Null values have been replaced"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8aa08063-6fa2-48de-82bd-a6c58ecebf90",
   "metadata": {},
   "source": [
    "## Step 3:Creating the windowed dataset for LSTM"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4b07f323-4f50-467e-9bb0-d0eefeb97b99",
   "metadata": {},
   "source": [
    "### 3a. Define function for the dataset creation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "347fb8a2",
   "metadata": {},
   "outputs": [],
   "source": [
    "def create_dataset(data, look_back=1):\n",
    "    X = [[data[i+j] for i in range(look_back)] for j in range(len(data)-look_back)]\n",
    "    X = np.array(X)\n",
    "    y = [data[i] for i in range(len(data)-look_back)]\n",
    "    y = np.array(y)\n",
    "    print(f'X shape = {X.shape} and y shape = {y.shape}')\n",
    "    return X,y"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "392d76bf-6a92-4832-b696-35299be2d963",
   "metadata": {},
   "source": [
    "### 3b. Scale data and create windowed dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "9d5755d5",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "X shape = (35039, 25, 1) and y shape = (35039, 1)\n"
     ]
    }
   ],
   "source": [
    "look_back = 25 # Changing may improve result\n",
    "\n",
    "data = imputed_df['total load actual'].astype('float32') \n",
    "data = np.reshape(data, (-1, 1)) # reshape to one feature; required for the models\n",
    "\n",
    "scaler = MinMaxScaler(feature_range=(0, 1)) # Min Max scaler\n",
    "data = scaler.fit_transform(data) # fit and transform the dataset\n",
    "\n",
    "X,y = create_dataset(data, look_back) # windowed dataset"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e40388ba-1998-4a1e-9172-4acbf7dcf5d1",
   "metadata": {},
   "source": [
    "### 3c. Split data into train, val and test subsets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "beb7c3db-04e3-48d4-8fa7-c09668df95b8",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training dataset shape: (22424, 25, 1)\n"
     ]
    }
   ],
   "source": [
    "X1, X_test, y1, y_test = train_test_split(X, y, test_size=0.2, random_state = 42)\n",
    "X_train, X_val, y_train, y_val = train_test_split(X1, y1, test_size=0.2, random_state = 42)\n",
    "\n",
    "print(f'Training dataset shape: {X_train.shape}')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e43fb213-d7a4-43e1-89fd-707881b152d6",
   "metadata": {},
   "source": [
    "### 3d. Define and train LSTM Model with the windowed dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "4ba7a032",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/50\n",
      "45/45 - 8s - loss: 0.3009 - accuracy: 0.0000e+00 - val_loss: 0.1942 - val_accuracy: 0.0000e+00 - 8s/epoch - 175ms/step\n",
      "Epoch 2/50\n",
      "45/45 - 1s - loss: 0.1673 - accuracy: 0.0000e+00 - val_loss: 0.1512 - val_accuracy: 0.0000e+00 - 864ms/epoch - 19ms/step\n",
      "Epoch 3/50\n",
      "45/45 - 1s - loss: 0.1425 - accuracy: 0.0000e+00 - val_loss: 0.1329 - val_accuracy: 0.0000e+00 - 791ms/epoch - 18ms/step\n",
      "Epoch 4/50\n",
      "45/45 - 1s - loss: 0.1209 - accuracy: 0.0000e+00 - val_loss: 0.1046 - val_accuracy: 0.0000e+00 - 784ms/epoch - 17ms/step\n",
      "Epoch 5/50\n",
      "45/45 - 1s - loss: 0.0908 - accuracy: 0.0000e+00 - val_loss: 0.0799 - val_accuracy: 0.0000e+00 - 753ms/epoch - 17ms/step\n",
      "Epoch 6/50\n",
      "45/45 - 1s - loss: 0.0840 - accuracy: 0.0000e+00 - val_loss: 0.0732 - val_accuracy: 0.0000e+00 - 731ms/epoch - 16ms/step\n",
      "Epoch 7/50\n",
      "45/45 - 1s - loss: 0.0750 - accuracy: 0.0000e+00 - val_loss: 0.0676 - val_accuracy: 0.0000e+00 - 749ms/epoch - 17ms/step\n",
      "Epoch 8/50\n",
      "45/45 - 1s - loss: 0.0700 - accuracy: 0.0000e+00 - val_loss: 0.0627 - val_accuracy: 0.0000e+00 - 731ms/epoch - 16ms/step\n",
      "Epoch 9/50\n",
      "45/45 - 1s - loss: 0.0646 - accuracy: 0.0000e+00 - val_loss: 0.0974 - val_accuracy: 0.0000e+00 - 754ms/epoch - 17ms/step\n",
      "Epoch 10/50\n",
      "45/45 - 1s - loss: 0.0640 - accuracy: 0.0000e+00 - val_loss: 0.0558 - val_accuracy: 0.0000e+00 - 754ms/epoch - 17ms/step\n",
      "Epoch 11/50\n",
      "45/45 - 1s - loss: 0.0596 - accuracy: 0.0000e+00 - val_loss: 0.0588 - val_accuracy: 0.0000e+00 - 738ms/epoch - 16ms/step\n",
      "Epoch 12/50\n",
      "45/45 - 1s - loss: 0.0597 - accuracy: 0.0000e+00 - val_loss: 0.0527 - val_accuracy: 0.0000e+00 - 754ms/epoch - 17ms/step\n",
      "Epoch 13/50\n",
      "45/45 - 1s - loss: 0.0538 - accuracy: 0.0000e+00 - val_loss: 0.0497 - val_accuracy: 0.0000e+00 - 733ms/epoch - 16ms/step\n",
      "Epoch 14/50\n",
      "45/45 - 1s - loss: 0.0518 - accuracy: 0.0000e+00 - val_loss: 0.0455 - val_accuracy: 0.0000e+00 - 739ms/epoch - 16ms/step\n",
      "Epoch 15/50\n",
      "45/45 - 1s - loss: 0.0505 - accuracy: 0.0000e+00 - val_loss: 0.0502 - val_accuracy: 0.0000e+00 - 792ms/epoch - 18ms/step\n",
      "Epoch 16/50\n",
      "45/45 - 1s - loss: 0.0487 - accuracy: 0.0000e+00 - val_loss: 0.0487 - val_accuracy: 0.0000e+00 - 756ms/epoch - 17ms/step\n",
      "Epoch 17/50\n",
      "45/45 - 1s - loss: 0.0479 - accuracy: 0.0000e+00 - val_loss: 0.0464 - val_accuracy: 0.0000e+00 - 744ms/epoch - 17ms/step\n",
      "Epoch 18/50\n",
      "45/45 - 1s - loss: 0.0448 - accuracy: 0.0000e+00 - val_loss: 0.0453 - val_accuracy: 0.0000e+00 - 751ms/epoch - 17ms/step\n",
      "Epoch 19/50\n",
      "45/45 - 1s - loss: 0.0472 - accuracy: 0.0000e+00 - val_loss: 0.0682 - val_accuracy: 0.0000e+00 - 740ms/epoch - 16ms/step\n",
      "Epoch 20/50\n",
      "45/45 - 1s - loss: 0.0454 - accuracy: 0.0000e+00 - val_loss: 0.0438 - val_accuracy: 0.0000e+00 - 747ms/epoch - 17ms/step\n",
      "Epoch 21/50\n",
      "45/45 - 1s - loss: 0.0452 - accuracy: 0.0000e+00 - val_loss: 0.0371 - val_accuracy: 0.0000e+00 - 758ms/epoch - 17ms/step\n",
      "Epoch 22/50\n",
      "45/45 - 1s - loss: 0.0414 - accuracy: 0.0000e+00 - val_loss: 0.0479 - val_accuracy: 0.0000e+00 - 805ms/epoch - 18ms/step\n",
      "Epoch 23/50\n",
      "45/45 - 1s - loss: 0.0411 - accuracy: 0.0000e+00 - val_loss: 0.0345 - val_accuracy: 0.0000e+00 - 816ms/epoch - 18ms/step\n",
      "Epoch 24/50\n",
      "45/45 - 1s - loss: 0.0462 - accuracy: 0.0000e+00 - val_loss: 0.0375 - val_accuracy: 0.0000e+00 - 772ms/epoch - 17ms/step\n",
      "Epoch 25/50\n",
      "45/45 - 1s - loss: 0.0389 - accuracy: 0.0000e+00 - val_loss: 0.0344 - val_accuracy: 0.0000e+00 - 790ms/epoch - 18ms/step\n",
      "Epoch 26/50\n",
      "45/45 - 1s - loss: 0.0417 - accuracy: 0.0000e+00 - val_loss: 0.0446 - val_accuracy: 0.0000e+00 - 742ms/epoch - 16ms/step\n",
      "Epoch 27/50\n",
      "45/45 - 1s - loss: 0.0402 - accuracy: 0.0000e+00 - val_loss: 0.0388 - val_accuracy: 0.0000e+00 - 731ms/epoch - 16ms/step\n",
      "Epoch 28/50\n",
      "45/45 - 1s - loss: 0.0417 - accuracy: 0.0000e+00 - val_loss: 0.0345 - val_accuracy: 0.0000e+00 - 732ms/epoch - 16ms/step\n",
      "Epoch 29/50\n",
      "45/45 - 1s - loss: 0.0405 - accuracy: 0.0000e+00 - val_loss: 0.0330 - val_accuracy: 0.0000e+00 - 726ms/epoch - 16ms/step\n",
      "Epoch 30/50\n",
      "45/45 - 1s - loss: 0.0359 - accuracy: 0.0000e+00 - val_loss: 0.0312 - val_accuracy: 0.0000e+00 - 742ms/epoch - 16ms/step\n",
      "Epoch 31/50\n",
      "45/45 - 1s - loss: 0.0366 - accuracy: 0.0000e+00 - val_loss: 0.0464 - val_accuracy: 0.0000e+00 - 723ms/epoch - 16ms/step\n",
      "Epoch 32/50\n",
      "45/45 - 1s - loss: 0.0416 - accuracy: 0.0000e+00 - val_loss: 0.0330 - val_accuracy: 0.0000e+00 - 726ms/epoch - 16ms/step\n",
      "Epoch 33/50\n",
      "45/45 - 1s - loss: 0.0350 - accuracy: 0.0000e+00 - val_loss: 0.0393 - val_accuracy: 0.0000e+00 - 746ms/epoch - 17ms/step\n",
      "Epoch 34/50\n",
      "45/45 - 1s - loss: 0.0374 - accuracy: 0.0000e+00 - val_loss: 0.0372 - val_accuracy: 0.0000e+00 - 726ms/epoch - 16ms/step\n",
      "Epoch 35/50\n",
      "45/45 - 1s - loss: 0.0377 - accuracy: 0.0000e+00 - val_loss: 0.0467 - val_accuracy: 0.0000e+00 - 732ms/epoch - 16ms/step\n",
      "Epoch 36/50\n",
      "45/45 - 1s - loss: 0.0334 - accuracy: 0.0000e+00 - val_loss: 0.0398 - val_accuracy: 0.0000e+00 - 740ms/epoch - 16ms/step\n",
      "Epoch 37/50\n",
      "45/45 - 1s - loss: 0.0369 - accuracy: 0.0000e+00 - val_loss: 0.0315 - val_accuracy: 0.0000e+00 - 755ms/epoch - 17ms/step\n",
      "Epoch 38/50\n",
      "45/45 - 1s - loss: 0.0331 - accuracy: 0.0000e+00 - val_loss: 0.0318 - val_accuracy: 0.0000e+00 - 742ms/epoch - 16ms/step\n",
      "Epoch 39/50\n",
      "45/45 - 1s - loss: 0.0336 - accuracy: 0.0000e+00 - val_loss: 0.0287 - val_accuracy: 0.0000e+00 - 720ms/epoch - 16ms/step\n",
      "Epoch 40/50\n",
      "45/45 - 1s - loss: 0.0484 - accuracy: 0.0000e+00 - val_loss: 0.0432 - val_accuracy: 0.0000e+00 - 743ms/epoch - 17ms/step\n",
      "Epoch 41/50\n",
      "45/45 - 1s - loss: 0.0419 - accuracy: 0.0000e+00 - val_loss: 0.0301 - val_accuracy: 0.0000e+00 - 744ms/epoch - 17ms/step\n",
      "Epoch 42/50\n",
      "45/45 - 1s - loss: 0.0323 - accuracy: 0.0000e+00 - val_loss: 0.0276 - val_accuracy: 0.0000e+00 - 742ms/epoch - 16ms/step\n",
      "Epoch 43/50\n",
      "45/45 - 1s - loss: 0.0324 - accuracy: 0.0000e+00 - val_loss: 0.0277 - val_accuracy: 0.0000e+00 - 784ms/epoch - 17ms/step\n",
      "Epoch 44/50\n",
      "45/45 - 1s - loss: 0.0363 - accuracy: 0.0000e+00 - val_loss: 0.0338 - val_accuracy: 0.0000e+00 - 792ms/epoch - 18ms/step\n",
      "Epoch 45/50\n",
      "45/45 - 1s - loss: 0.0358 - accuracy: 0.0000e+00 - val_loss: 0.0379 - val_accuracy: 0.0000e+00 - 798ms/epoch - 18ms/step\n",
      "Epoch 46/50\n",
      "45/45 - 1s - loss: 0.0317 - accuracy: 0.0000e+00 - val_loss: 0.0270 - val_accuracy: 0.0000e+00 - 799ms/epoch - 18ms/step\n",
      "Epoch 47/50\n",
      "45/45 - 1s - loss: 0.0332 - accuracy: 0.0000e+00 - val_loss: 0.0372 - val_accuracy: 0.0000e+00 - 740ms/epoch - 16ms/step\n",
      "Epoch 48/50\n",
      "45/45 - 1s - loss: 0.0318 - accuracy: 0.0000e+00 - val_loss: 0.0328 - val_accuracy: 0.0000e+00 - 741ms/epoch - 16ms/step\n",
      "Epoch 49/50\n",
      "45/45 - 1s - loss: 0.0324 - accuracy: 0.0000e+00 - val_loss: 0.0309 - val_accuracy: 0.0000e+00 - 772ms/epoch - 17ms/step\n",
      "Epoch 50/50\n",
      "45/45 - 1s - loss: 0.0320 - accuracy: 0.0000e+00 - val_loss: 0.0340 - val_accuracy: 0.0000e+00 - 768ms/epoch - 17ms/step\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA1cAAAIhCAYAAACizkCYAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjguMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy81sbWrAAAACXBIWXMAAA9hAAAPYQGoP6dpAACODElEQVR4nOzdd3RU1d7G8e+kJ0BCKAkgoZfQS+gIgjQBESwXbCCKvYLXK3IRC3ovlqtgA8urIBaMCiIqSm8CIi0gEhCkJEBCT0JIz5z3j00SQhJIn0zyfNY6a86cc+bMnhh1nuy9f9tmWZaFiIiIiIiIFImLoxsgIiIiIiJSHihciYiIiIiIFAOFKxERERERkWKgcCUiIiIiIlIMFK5ERERERESKgcKViIiIiIhIMVC4EhERERERKQYKVyIiIiIiIsVA4UpERERERKQYKFyJiEiBzJkzB5vNxpYtWxzdlDIn42eT17Z69WqHtu/QoUPYbDb+97//ObQdIiLllZujGyAiIlLezJ49m+Dg4BzHW7Zs6YDWiIhIaVG4EhERKWatW7emU6dOjm6GiIiUMg0LFBGREvHrr7/Sr18/qlSpgo+PDz169OCnn37Kdk1CQgJPPfUUDRs2xMvLi2rVqtGpUyfmzZuXec2BAwe49dZbqVOnDp6engQGBtKvXz/CwsLyfO8ZM2Zgs9nYv39/jnMTJ07Ew8ODU6dOAbB9+3auv/56AgIC8PT0pE6dOgwdOpQjR44Uzw8iDzabjUcffZQPPviAZs2a4enpScuWLfnqq69yXLtr1y6GDx+Ov78/Xl5etG/fnk8//TTHdTExMfzzn/+kUaNGeHp6EhAQwJAhQ9izZ0+Oa998800aNmxI5cqV6d69O7/99luJfE4RkYpEPVciIlLs1qxZw4ABA2jbti0ff/wxnp6ezJw5k2HDhjFv3jxGjRoFwJNPPslnn33Gyy+/TIcOHTh//jy7du3i9OnTmfcaMmQI6enpvPbaa9SrV49Tp06xYcMGYmJi8nz/O++8k4kTJzJnzhxefvnlzOPp6el8/vnnDBs2jBo1anD+/HkGDBhAw4YNee+99wgMDCQ6OppVq1Zx7ty5Qn/+9PR00tLSsh2z2Wy4urpmO7Zo0SJWrVrF1KlTqVSpEjNnzuS2227Dzc2NW265BYC9e/fSo0cPAgICePvtt6levTqff/45Y8eO5fjx4zz99NMAnDt3jquvvppDhw4xceJEunbtSnx8PGvXriUqKirbMMX33nuP4OBgZsyYAcCUKVMYMmQIBw8exM/Pr9CfW0SkwrNEREQKYPbs2RZgbd68Oc9runXrZgUEBFjnzp3LPJaWlma1bt3aqlu3rmW32y3LsqzWrVtbI0aMyPM+p06dsgBrxowZBW7nTTfdZNWtW9dKT0/PPLZ48WILsH744QfLsixry5YtFmAtXLiwwPfPTcbPJrfN1dU127WA5e3tbUVHR2ceS0tLs4KDg60mTZpkHrv11lstT09PKyIiItvrBw8ebPn4+FgxMTGWZVnW1KlTLcBatmxZnu07ePCgBVht2rSx0tLSMo///vvvFmDNmzevSJ9fRKSi07BAEREpVufPn2fTpk3ccsstVK5cOfO4q6sro0eP5siRI+zduxeALl268PPPP/PMM8+wevVqEhMTs92rWrVqNG7cmNdff50333yT7du3Y7fb89WOu+++myNHjrB8+fLMY7Nnz6ZWrVoMHjwYgCZNmuDv78/EiRN5//332b17d1E/PgBz585l8+bN2bZNmzbluK5fv34EBgZmPnd1dWXUqFHs378/c1jiypUr6devH0FBQdleO3bsWBISEti4cSMAP//8M82aNaN///5XbN/QoUOz9aK1bdsWgMOHDxf8w4qISCaFKxERKVZnz57Fsixq166d41ydOnUAMof9vf3220ycOJGFCxfSt29fqlWrxogRI9i3bx9ghtKtWLGCQYMG8dprr9GxY0dq1qzJ448/fsVhe4MHD6Z27drMnj07s12LFi1izJgxmcHCz8+PNWvW0L59e/7973/TqlUr6tSpw/PPP09qamqhfwYtWrSgU6dO2baQkJAc19WqVSvPYxk/o9OnT+frZ3ny5Enq1q2br/ZVr14923NPT0+AHOFWREQKRuFKRESKlb+/Py4uLkRFReU4d+zYMQBq1KgBQKVKlXjxxRfZs2cP0dHRzJo1i99++41hw4ZlvqZ+/fp8/PHHREdHs3fvXiZMmMDMmTP517/+ddl2ZPSULVy4kJiYGL788kuSk5O5++67s13Xpk0bvvrqK06fPk1YWBijRo1i6tSpvPHGG0X9UVxRdHR0nscyAlD16tXz9bOsWbNmiRfhEBGRy1O4EhGRYlWpUiW6du3KggULsvWE2O12Pv/8c+rWrUuzZs1yvC4wMJCxY8dy2223sXfvXhISEnJc06xZM5599lnatGnDtm3brtiWu+++m6SkJObNm8ecOXPo3r17rutPgekla9euHdOnT6dq1ar5un9RrVixguPHj2c+T09PJzQ0lMaNG2f2QvXr14+VK1dmhqkMc+fOxcfHh27dugGmp+6vv/5i5cqVJd5uERHJnaoFiohIoaxcuZJDhw7lOD5kyBCmTZvGgAED6Nu3L0899RQeHh7MnDmTXbt2MW/ePGw2GwBdu3bl+uuvp23btvj7+xMeHs5nn31G9+7d8fHxYefOnTz66KP84x//oGnTpnh4eLBy5Up27tzJM888c8U2BgcH0717d6ZNm0ZkZCQffvhhtvM//vgjM2fOZMSIETRq1AjLsliwYAExMTEMGDAg87p+/fqxZs2aHBUA87Jr165cr23cuDE1a9bMfF6jRg2uvfZapkyZklktcM+ePdnKsT///PP8+OOP9O3bl+eee45q1arxxRdf8NNPP/Haa69lVvcbP348oaGhDB8+nGeeeYYuXbqQmJjImjVruP766+nbt2++2i4iIkXg6IoaIiLiXC5XEQ+wDh48aFmWZa1bt8669tprrUqVKlne3t5Wt27dMqv0ZXjmmWesTp06Wf7+/panp6fVqFEja8KECdapU6csy7Ks48ePW2PHjrWCg4OtSpUqWZUrV7batm1rTZ8+PVu1u8v58MMPMyvzxcbGZju3Z88e67bbbrMaN25seXt7W35+flaXLl2sOXPmZLvummuusfLzv8wr/Ww++uijzGsB65FHHrFmzpxpNW7c2HJ3d7eCg4OtL774Isd9//jjD2vYsGGWn5+f5eHhYbVr186aPXt2juvOnj1rPfHEE1a9evUsd3d3KyAgwBo6dKi1Z88ey7KyqgW+/vrrOV4LWM8///wVP6OIiOTNZlmWVcp5TkREpMKz2Ww88sgjvPvuu45uioiIFBPNuRIRERERESkGClciIiIiIiLFQAUtREREHECj8kVEyh/1XImIiIiIiBQDhSsREREREZFioHAlIiIiIiJSDDTnKhd2u51jx45RpUqVzIUuRURERESk4rEsi3PnzlGnTh1cXC7fN6VwlYtjx44RFBTk6GaIiIiIiEgZERkZSd26dS97jcJVLqpUqQKYH6Cvr6+DWyMiIiIiIo4SFxdHUFBQZka4HIWrXGQMBfT19VW4EhERERGRfE0XUkELERERERGRYqBwJSIiIiIiUgwUrkRERERERIqB5lyJiIiIiFOwLIu0tDTS09Md3RQpZ9zd3XF1dS3yfRSuRERERKTMS0lJISoqioSEBEc3Rcohm81G3bp1qVy5cpHuo3AlIiIiImWa3W7n4MGDuLq6UqdOHTw8PPJVuU0kPyzL4uTJkxw5coSmTZsWqQdL4UpEREREyrSUlBTsdjtBQUH4+Pg4ujlSDtWsWZNDhw6RmppapHClghYiIiIi4hRcXPTVVUpGcfWE6jdURERERESkGChciYiIiIiIFAOFKxERERERJ9GnTx/Gjx+f7+sPHTqEzWYjLCysxNokWRSuRERERESKmc1mu+w2duzYQt13wYIFvPTSS/m+PigoiKioKFq3bl2o98svhThD1QJFRERERIpZVFRU5n5oaCjPPfcce/fuzTzm7e2d7frU1FTc3d2veN9q1aoVqB2urq7UqlWrQK+RwlPPlYiIiIg4FcuC8+cds1lW/tpYq1atzM3Pzw+bzZb5PCkpiapVq/L111/Tp08fvLy8+Pzzzzl9+jS33XYbdevWxcfHhzZt2jBv3rxs9710WGCDBg3473//yz333EOVKlWoV68eH374Yeb5S3uUVq9ejc1mY8WKFXTq1AkfHx969OiRLfgBvPzyywQEBFClShXuvfdennnmGdq3b1+Yf1wAJCcn8/jjjxMQEICXlxdXX301mzdvzjx/9uxZ7rjjDmrWrIm3tzdNmzZl9uzZgCnF/+ijj1K7dm28vLxo0KAB06ZNK3RbSpLClYiIiIg4lYQEqFzZMVtCQvF9jokTJ/L4448THh7OoEGDSEpKIiQkhB9//JFdu3Zx//33M3r0aDZt2nTZ+7zxxht06tSJ7du38/DDD/PQQw+xZ8+ey75m8uTJvPHGG2zZsgU3NzfuueeezHNffPEF//nPf3j11VfZunUr9erVY9asWUX6rE8//TTz58/n008/Zdu2bTRp0oRBgwZx5swZAKZMmcLu3bv5+eefCQ8PZ9asWdSoUQOAt99+m0WLFvH111+zd+9ePv/8cxo0aFCk9pQUh4ermTNn0rBhQ7y8vAgJCWHdunV5Xvvrr7/Ss2dPqlevjre3N8HBwUyfPj3HdfPnz6dly5Z4enrSsmVLvvvuu5L8CCIiIiIiBTZ+/HhuuukmGjZsSJ06dbjqqqt46qmnaN++PY0aNeKxxx5j0KBBfPPNN5e9z5AhQ3j44Ydp0qQJEydOpEaNGqxevfqyr/nPf/7DNddcQ8uWLXnmmWfYsGEDSUlJALzzzjuMGzeOu+++m2bNmvHcc8/Rpk2bQn/O8+fPM2vWLF5//XUGDx5My5Yt+eijj/D29ubjjz8GICIigg4dOtCpUycaNGhA//79GTZsWOa5pk2bcvXVV1O/fn2uvvpqbrvttkK3pyQ5NFyFhoYyfvx4Jk+ezPbt2+nVqxeDBw8mIiIi1+srVarEo48+ytq1awkPD+fZZ5/l2Wefzdb1uXHjRkaNGsXo0aPZsWMHo0ePZuTIkVdM/GVVeDjMnw87dji6JSIiIiJlg48PxMc7ZvPxKb7P0alTp2zP09PT+c9//kPbtm2pXr06lStXZunSpXl+N87Qtm3bzP2M4YcnTpzI92tq164NkPmavXv30qVLl2zXX/q8IP7++29SU1Pp2bNn5jF3d3e6dOlCeHg4AA899BBfffUV7du35+mnn2bDhg2Z144dO5awsDCaN2/O448/ztKlSwvdlpLm0HD15ptvMm7cOO69915atGjBjBkzCAoKyrPbsUOHDtx22220atWKBg0acOeddzJo0KBsvV0zZsxgwIABTJo0ieDgYCZNmkS/fv2YMWNGKX2q4vXRR3DLLfDll45uiYiIiEjZYLNBpUqO2Wy24vsclSpVyvb8jTfeYPr06Tz99NOsXLmSsLAwBg0aREpKymXvc2khDJvNht1uz/drbBc+1MWvsV3yQa38TjbLRcZrc7tnxrHBgwdz+PBhxo8fz7Fjx+jXrx9PPfUUAB07duTgwYO89NJLJCYmMnLkSG655ZZCt6ckOSxcpaSksHXrVgYOHJjt+MCBA7Ml1cvZvn07GzZs4Jprrsk8tnHjxhz3HDRo0GXvmZycTFxcXLatrMgoCHP2rGPbISIiIiIla926dQwfPpw777yTdu3a0ahRI/bt21fq7WjevDm///57tmNbtmwp9P2aNGmCh4cHv/76a+ax1NRUtmzZQosWLTKP1axZk7Fjx/L5558zY8aMbKPTfH19GTVqFB999BGhoaHMnz8/c75WWeKwUuynTp0iPT2dwMDAbMcDAwOJjo6+7Gvr1q3LyZMnSUtL44UXXuDee+/NPBcdHV3ge06bNo0XX3yxEJ+i5GWEqzL4uyMiIiIixahJkybMnz+fDRs24O/vz5tvvkl0dHS2AFIaHnvsMe677z46depEjx49CA0NZefOnTRq1OiKr7206iBAy5Yteeihh/jXv/5FtWrVqFevHq+99hoJCQmMGzcOgOeee46QkBBatWpFcnIyP/74Y+bnnj59OrVr16Z9+/a4uLjwzTffUKtWLapWrVqsn7s4OHydq8t1D+Zl3bp1xMfH89tvv/HMM8/QpEmTbJPaCnrPSZMm8eSTT2Y+j4uLIygoqCAfo8QoXImIiIhUDFOmTOHgwYMMGjQIHx8f7r//fkaMGEFsbGyptuOOO+7gwIEDPPXUUyQlJTFy5EjGjh2bozcrN7feemuOYwcPHuSVV17BbrczevRozp07R6dOnViyZAn+/v4AeHh4MGnSJA4dOoS3tze9evXiq6++AqBy5cq8+uqr7Nu3D1dXVzp37szixYtxcXF4bb4cbFZRBlAWQUpKCj4+PnzzzTfceOONmcefeOIJwsLCWLNmTb7u8/LLL/PZZ59lpuR69eoxYcIEJkyYkHnN9OnTmTFjBocPH87XPePi4vDz8yM2NhZfX98CfKrit3QpDBoE7dpBBV/wWkRERCqopKQkDh48mFlhWkrfgAEDqFWrFp999pmjm1IiLvc7VpBs4LC45+HhQUhICMuWLct2fNmyZfTo0SPf97Esi+Tk5Mzn3bt3z3HPpUuXFuieZYl6rkRERESkNCUkJPDmm2/y559/smfPHp5//nmWL1/OXXfd5eimlXkOHRb45JNPMnr0aDp16kT37t358MMPiYiI4MEHHwTMcL2jR48yd+5cAN577z3q1atHcHAwYNa9+t///sdjjz2Wec8nnniC3r178+qrrzJ8+HC+//57li9fnm0CnTNRuBIRERGR0mSz2Vi8eDEvv/wyycnJNG/enPnz59O/f39HN63Mc2i4GjVqFKdPn2bq1KlERUXRunVrFi9eTP369QGIiorKVtffbrczadIkDh48iJubG40bN+aVV17hgQceyLymR48efPXVVzz77LNMmTKFxo0bExoaSteuXUv98xWHjHB1/jwkJ4Onp2PbIyIiIiLlm7e3N8uXL3d0M5ySw+ZclWVlac6V3Q7u7uYxKgpq1XJoc0RERERKneZcSUlz+jlXkj8uLpBRZVJDA0VEREREyi6FKyegeVciIiIiImWfwpUTULgSERERESn7FK6cgMKViIiIiEjZp3DlBBSuRERERETKPoUrJ6BwJSIiIlIx9enTh/Hjx2c+b9CgATNmzLjsa2w2GwsXLizyexfXfSoShSsnoHAlIiIi4lyGDRuW56K7GzduxGazsW3btgLfd/Pmzdx///1FbV42L7zwAu3bt89xPCoqisGDBxfre11qzpw5VM0ojV0OKFw5AYUrEREREecybtw4Vq5cyeHDh3Oc++STT2jfvj0dO3Ys8H1r1qyJj49PcTTximrVqoWnp2epvFd5oXDlBBSuRERERC5iWZB23jGbZeWriddffz0BAQHMmTMn2/GEhARCQ0MZN24cp0+f5rbbbqNu3br4+PjQpk0b5s2bd9n7XjoscN++ffTu3RsvLy9atmzJsmXLcrxm4sSJNGvWDB8fHxo1asSUKVNITU0FTM/Riy++yI4dO7DZbNhstsw2Xzos8I8//uDaa6/F29ub6tWrc//99xMfH595fuzYsYwYMYL//e9/1K5dm+rVq/PII49kvldhREREMHz4cCpXroyvry8jR47k+PHjmed37NhB3759qVKlCr6+voSEhLBlyxYADh8+zLBhw/D396dSpUq0atWKxYsXF7ot+eFWoneXYqFwJSIiInKR9AT4urJj3ntkPLhVuuJlbm5ujBkzhjlz5vDcc89hs9kA+Oabb0hJSeGOO+4gISGBkJAQJk6ciK+vLz/99BOjR4+mUaNGdO3a9YrvYbfbuemmm6hRowa//fYbcXFx2eZnZahSpQpz5syhTp06/PHHH9x3331UqVKFp59+mlGjRrFr1y5++eUXli9fDoCfn1+OeyQkJHDdddfRrVs3Nm/ezIkTJ7j33nt59NFHswXIVatWUbt2bVatWsX+/fsZNWoU7du357777rvi57mUZVmMGDGCSpUqsWbNGtLS0nj44YcZNWoUq1evBuCOO+6gQ4cOzJo1C1dXV8LCwnB3dwfgkUceISUlhbVr11KpUiV2795N5col+3ujcOUEFK5EREREnM8999zD66+/zurVq+nbty9ghgTedNNN+Pv74+/vz1NPPZV5/WOPPcYvv/zCN998k69wtXz5csLDwzl06BB169YF4L///W+OeVLPPvts5n6DBg345z//SWhoKE8//TTe3t5UrlwZNzc3atWqled7ffHFFyQmJjJ37lwqVTLh8t1332XYsGG8+uqrBAYGAuDv78+7776Lq6srwcHBDB06lBUrVhQqXC1fvpydO3dy8OBBgoKCAPjss89o1aoVmzdvpnPnzkRERPCvf/2L4OBgAJo2bZr5+oiICG6++WbatGkDQKNGjQrchoJSuHICClciIiIiF3H1MT1IjnrvfAoODqZHjx588skn9O3bl7///pt169axdOlSANLT03nllVcIDQ3l6NGjJCcnk5ycnBleriQ8PJx69eplBiuA7t2757ju22+/ZcaMGezfv5/4+HjS0tLw9fXN9+fIeK927dpla1vPnj2x2+3s3bs3M1y1atUKV1fXzGtq167NH3/8UaD3uvg9g4KCMoMVQMuWLalatSrh4eF07tyZJ598knvvvZfPPvuM/v37849//IPGjRsD8Pjjj/PQQw+xdOlS+vfvz80330zbtm0L1Zb80pwrJ5ARrmJjIS3NsW0RERERcTibzQzNc8R2YXhffo0bN4758+cTFxfH7NmzqV+/Pv369QPgjTfeYPr06Tz99NOsXLmSsLAwBg0aREpKSr7ubeUy/8t2Sft+++03br31VgYPHsyPP/7I9u3bmTx5cr7f4+L3uvTeub1nxpC8i8/Z7fYCvdeV3vPi4y+88AJ//vknQ4cOZeXKlbRs2ZLvvvsOgHvvvZcDBw4wevRo/vjjDzp16sQ777xTqLbkl8KVE/D3z9qPiXFYM0RERESkgEaOHImrqytffvkln376KXfffXdmMFi3bh3Dhw/nzjvvpF27djRq1Ih9+/bl+94tW7YkIiKCY8eOZR7buHFjtmvWr19P/fr1mTx5Mp06daJp06Y5Khh6eHiQnp5+xfcKCwvj/Pnz2e7t4uJCs2bN8t3mgsj4fJGRkZnHdu/eTWxsLC1atMg81qxZMyZMmMDSpUu56aabmD17dua5oKAgHnzwQRYsWMA///lPPvrooxJpawaFKyfg5gYZPbcaGigiIiLiPCpXrsyoUaP497//zbFjxxg7dmzmuSZNmrBs2TI2bNhAeHg4DzzwANHR0fm+d//+/WnevDljxoxhx44drFu3jsmTJ2e7pkmTJkRERPDVV1/x999/8/bbb2f27GRo0KABBw8eJCwsjFOnTpGcnJzjve644w68vLy466672LVrF6tWreKxxx5j9OjRmUMCCys9PZ2wsLBs2+7du+nfvz9t27bljjvuYNu2bfz++++MGTOGa665hk6dOpGYmMijjz7K6tWrOXz4MOvXr2fz5s2ZwWv8+PEsWbKEgwcPsm3bNlauXJktlJUEhSsnoXlXIiIiIs5p3LhxnD17lv79+1OvXr3M41OmTKFjx44MGjSIPn36UKtWLUaMGJHv+7q4uPDdd9+RnJxMly5duPfee/nPf/6T7Zrhw4czYcIEHn30Udq3b8+GDRuYMmVKtmtuvvlmrrvuOvr27UvNmjVzLQfv4+PDkiVLOHPmDJ07d+aWW26hX79+vPvuuwX7YeQiPj6eDh06ZNuGDBmSWQre39+f3r17079/fxo1akRoaCgArq6unD59mjFjxtCsWTNGjhzJ4MGDefHFFwET2h555BFatGjBddddR/PmzZk5c2aR23s5Niu3wZoVXFxcHH5+fsTGxhZ4sl9JCQmBbdvgp59gyBBHt0ZERESk9CQlJXHw4EEaNmyIl5eXo5sj5dDlfscKkg3Uc+Uk1HMlIiIiIlK2KVw5CYUrEREREZGyTeHKSWSEq7NnHdsOERERERHJncKVk1DPlYiIiIhI2aZw5SQUrkRERKSiUx02KSnF9bulcOUkFK5ERESkonJ3dwcgISHBwS2R8iolJQUw5d2Lwq04GiMlT+FKREREKipXV1eqVq3KiRMnALPmks1mc3CrpLyw2+2cPHkSHx8f3NyKFo8UrpyEwpWIiIhUZLVq1QLIDFgixcnFxYV69eoVObQrXDkJhSsRERGpyGw2G7Vr1yYgIIDU1FRHN0fKGQ8PD1xcij5jSuHKSVwcrux2KIZ/9iIiIiJOx9XVtcjzYkRKir6iOwl/f/Not8O5c45ti4iIiIiI5KRw5SS8vMDb2+xraKCIiIiISNmjcOVENO9KRERERKTsUrhyIgpXIiIiIiJll8KVE1G4EhEREREpuxSunIjClYiIiIhI2aVw5UQUrkREREREyi6FKyeicCUiIiIiUnYpXDkRhSsRERERkbJL4cqJKFyJiIiIiJRdCldOROFKRERERKTsUrhyIgpXIiIiIiJll8KVE1G4EhEREREpuxSunMjF4cqyHNsWERERERHJTuHKiWSEq5QUSEhwbFtERERERCQ7hSsnUqkSuLubfQ0NFBEREREpWxSunIjNpnlXIiIiIiJllcKVk1G4EhEREREpmxSunIzClYiIiIhI2aRw5WQywtXZs45th4iIiIiIZKdw5WTUcyUiIiIiUjYpXDkZhSsRERERkbJJ4crJKFyJiIiIiJRNCldORuFKRERERKRsUrhyMgpXIiIiIiJlk8KVk1G4EhEREREpmxSunIzClYiIiIhI2aRw5WQUrkREREREyiaFKyeTEa7On4fkZMe2RUREREREsihcORlfX7DZzP7Zs45ti4iIiIiIZFG4cjIuLuDvb/Y1NFBEREREpOxQuHJCmnclIiIiIlL2KFw5IYUrEREREZGyR+HKCSlciYiIiIiUPQpXTkjhSkRERESk7FG4ckIKVyIiIiIiZY/ClRNSuBIRERERKXsUrpyQwpWIiIiISNmjcOWEFK5ERERERMoeh4ermTNn0rBhQ7y8vAgJCWHdunV5XrtgwQIGDBhAzZo18fX1pXv37ixZsiTbNXPmzMFms+XYkpKSSvqjlBqFKxERERGRsseh4So0NJTx48czefJktm/fTq9evRg8eDARERG5Xr927VoGDBjA4sWL2bp1K3379mXYsGFs374923W+vr5ERUVl27y8vErjI5UKhSsRERERkbLHZlmW5ag379q1Kx07dmTWrFmZx1q0aMGIESOYNm1avu7RqlUrRo0axXPPPQeYnqvx48cTExNT6HbFxcXh5+dHbGwsvr6+hb5PSdm7F4KDwc8PivAxRURERETkCgqSDRzWc5WSksLWrVsZOHBgtuMDBw5kw4YN+bqH3W7n3LlzVMvoyrkgPj6e+vXrU7duXa6//vocPVuXSk5OJi4uLttWlmV83NhYSEtzbFtERERERMRwWLg6deoU6enpBAYGZjseGBhIdHR0vu7xxhtvcP78eUaOHJl5LDg4mDlz5rBo0SLmzZuHl5cXPXv2ZN++fXneZ9q0afj5+WVuQUFBhftQpcTfP2tfPVciIiIiImWDwwta2Gy2bM8ty8pxLDfz5s3jhRdeIDQ0lICAgMzj3bp1484776Rdu3b06tWLr7/+mmbNmvHOO+/kea9JkyYRGxubuUVGRhb+A5UCNzfI6JHUvCsRERERkbLBzVFvXKNGDVxdXXP0Up04cSJHb9alQkNDGTduHN988w39+/e/7LUuLi507tz5sj1Xnp6eeHp65r/xZUC1ahAXp3AlIiIiIlJWOKznysPDg5CQEJYtW5bt+LJly+jRo0eer5s3bx5jx47lyy+/ZOjQoVd8H8uyCAsLo3bt2kVuc1mSMe/q7FnHtkNERERERAyH9VwBPPnkk4wePZpOnTrRvXt3PvzwQyIiInjwwQcBM1zv6NGjzJ07FzDBasyYMbz11lt069Yts9fL29sbPz8/AF588UW6detG06ZNiYuL4+233yYsLIz33nvPMR+yhKgcu4iIiIhI2eLQcDVq1ChOnz7N1KlTiYqKonXr1ixevJj69esDEBUVlW3Nqw8++IC0tDQeeeQRHnnkkczjd911F3PmzAEgJiaG+++/n+joaPz8/OjQoQNr166lS5cupfrZSprClYiIiIhI2eLQda7KqrK+zhXAQw/B++/DCy/A8887ujUiIiIiIuWTU6xzJUWjnisRERERkbJF4cpJKVyJiIiIiJQtCldOSuFKRERERKRsUbhyUgpXIiIiIiJli8KVk1K4EhEREREpWxSunJTClYiIiIhI2aJw5aQuDld2u2PbIiIiIiIiCldOy9/fPNrtcO6cY9siIiIiIiIKV07Lywu8vc2+hgaKiIiIiDiewpUT07wrEREREZGyQ+HKiSlciYiIiIiUHQpXTkzhSkRERESk7FC4cmIKVyIiIiIiZYfClRNTuBIRERERKTsUrpyYwpWIiIiISNmhcOXEFK5ERERERMoOhSsnpnAlIiIiIlJ2KFw5MYUrEREREZGyQ+HKiSlciYiIiIiUHQpXTkzhSkRERESk7FC4cmIXhyvLcmxbREREREQqOoUrJ5YRrlJSICHBsW0REREREanoFK6cWKVK4O5u9jU0UERERETEsRSunJjNltV7dfasY9siIiIiIlLRKVw5ORW1EBEREREpGxSunJzClYiIiIhI2aBw5eQUrkREREREygaFKyencCUiIiIiUjYoXDk5hSsRERERkbJB4crJKVyJiIiIiJQNCldOTuFKRERERKRsULhycgpXIiIiIiJlg8KVk1O4EhEREREpGxSunJzClYiIiIhI2aBw5eQUrkREREREygaFKyfn728ez5+H5GTHtkVEREREpCJTuHJyfn5gs5n9s2cd2xYRERERkYpM4crJubhk9V5paKCIiIiIiOMoXJUDmnclIiIiIuJ4ClflgMKViIiIiIjjKVyVAwpXIiIiIiKOp3BVDihciYiIiIg4nsJVOaBwJSIiIiLieApX5YDClYiIiIiI4ylclQMKVyIiIiIijqdwVQ4oXImIiIiIOJ7CVTmgcCUiIiIi4ngKV+WAwpWIiIiIiOMpXJUDClciIiIiIo6ncFUOZISr2FhIT3dsW0REREREKiqFq3LA3z9rPybGYc0QEREREanQFK7KATc38PU1+xoaKCIiIiLiGApX5YTmXYmIiIiIOJbCVTmhcCUiIiIi4lgKV+WEwpWIiIiIiGMpXJUTClciIiIiIo6lcFVOKFyJiIiIiDiWwlU5oXAlIiIiIuJYClflhMKViIiIiIhjKVyVEwpXIiIiIiKOpXBVTihciYiIiIg4lsJVOaFwJSIiIiLiWApX5YTClYiIiIiIYylclRP+/ubxzBmw2x3bFhERERGRikjhqpzICFd2O5w759i2iIiIiIhURApX5YS3t9lAQwNFRERERBzB4eFq5syZNGzYEC8vL0JCQli3bl2e1y5YsIABAwZQs2ZNfH196d69O0uWLMlx3fz582nZsiWenp60bNmS7777riQ/QpmheVciIiIiIo7j0HAVGhrK+PHjmTx5Mtu3b6dXr14MHjyYiIiIXK9fu3YtAwYMYPHixWzdupW+ffsybNgwtm/fnnnNxo0bGTVqFKNHj2bHjh2MHj2akSNHsmnTptL6WA6jcCUiIiIi4jg2y7IsR715165d6dixI7Nmzco81qJFC0aMGMG0adPydY9WrVoxatQonnvuOQBGjRpFXFwcP//8c+Y11113Hf7+/sybNy/XeyQnJ5OcnJz5PC4ujqCgIGJjY/H19S3MR3OIPn1gzRr46isYNcrRrRERERERcX5xcXH4+fnlKxs4rOcqJSWFrVu3MnDgwGzHBw4cyIYNG/J1D7vdzrlz56iW0WWD6bm69J6DBg267D2nTZuGn59f5hYUFFSAT1J2qOdKRERERMRxHBauTp06RXp6OoGBgdmOBwYGEh0dna97vPHGG5w/f56RI0dmHouOji7wPSdNmkRsbGzmFhkZWYBPUnYoXImIiIiIOI6boxtgs9myPbcsK8ex3MybN48XXniB77//noCAgCLd09PTE09PzwK0umxSuBIRERERcRyHhasaNWrg6uqao0fpxIkTOXqeLhUaGsq4ceP45ptv6N+/f7ZztWrVKtQ9ywOFKxERERERx3HYsEAPDw9CQkJYtmxZtuPLli2jR48eeb5u3rx5jB07li+//JKhQ4fmON+9e/cc91y6dOll71leKFyJiIiIiDiOQ4cFPvnkk4wePZpOnTrRvXt3PvzwQyIiInjwwQcBMxfq6NGjzJ07FzDBasyYMbz11lt069Yts4fK29sbPz8/AJ544gl69+7Nq6++yvDhw/n+++9Zvnw5v/76q2M+ZClSuBIRERERcRyHrnM1atQoZsyYwdSpU2nfvj1r165l8eLF1K9fH4CoqKhsa1598MEHpKWl8cgjj1C7du3M7Yknnsi8pkePHnz11VfMnj2btm3bMmfOHEJDQ+natWupf77SpnAlIiIiIuI4Dl3nqqwqSC37siQsDDp0gFq1ICrK0a0REREREXF+TrHOleTTXzPhly7w9ydXvDSj5+rsWVBkFhEREREpXQpXZV3iMTizGU5eec5YRrhKTobExBJul4iIiIiIZKNwVdbV6GYeT/12xUsrVQJ3d7OveVciIiIiIqVL4aqsq36hEEdcOKTEXPZSm01FLUREREREHEXhqqzzqgmVG5n905uveLnClYiIiIiIYyhcOYPq+R8aqHAlIiIiIuIYClfOIGPe1WmFKxERERGRskrhyhlkhqtNV6yxrnAlIiIiIuIYClfOoGo7cPGE5NMQ//dlL1W4EhERERFxDIUrZ+DqAdU6mv0rzLtSuBIRERERcQyFK2eRz6IWClciIiIiIo6hcOUs8lnUQuFKRERERMQxFK6cRY0Liwmf3QFpiXlepnAlIiIiIuIYClfOwqceeNUCKw3ObsvzMoUrERERERHHULhyFjZb1tDAy8y78vc3jwpXIiIiIiKlS+HKmeQjXGX0XJ0/D8nJpdAmEREREREBFK6cS/UL865Ob8rzEj8/08kFcPZsKbRJREREREQAhSvnUq0T2FwgIRISjuZ6iYuLhgaKiIiIiDiCwpUzca8Mfm3M/mV6r1TUQkRERESk9ClcOZsCzLtSuBIRERERKT0KV84mH/OuFK5EREREREqfwpWzyei5Or0Z7Gm5XqJwJSIiIiJS+hSunI1vc3D3g/REiPkj10sUrkRERERESp/ClbOxuVw0NDD3eVcKVyIiIiIipU/hyhnVuBCuTuU+70rhSkRERESk9ClcOaPqGfOu1HMlIiIiIlJWKFw5o4yeq7i9kJwzQWWEq7NnS7FNIiIiIiIVnMKVM/KsDlWamv3Tv+c4rZ4rEREREZHSp3DlrC6z3pXClYiIiIhI6VO4clYZ612dyjnvKiNcxcRAenrpNUlEREREpCJTuHJWmYsJbwLLnu2Uv3/WfkxM6TVJRERERKQiU7hyVlXbgqsXpJyFc/uynXJzA19fs6+hgSIiIiIipUPhylm5uEO1TmY/l/WuNO9KRERERKR0KVw5s8yiFnnPu1K4EhEREREpHQpXziwfRS0UrkRERERESofClTPLCFcxOyEtIdsphSsRERERkdJVqHAVGRnJkSNHMp///vvvjB8/ng8//LDYGib54FMXvK8CKx3ObM12SuFKRERERKR0FSpc3X777axatQqA6OhoBgwYwO+//86///1vpk6dWqwNlCuocWHe1SVDAxWuRERERERKV6HC1a5du+jSpQsAX3/9Na1bt2bDhg18+eWXzJkzpzjbJ1dSPWO9K4UrERERERFHKlS4Sk1NxdPTE4Dly5dzww03ABAcHExUVFTxtU6uLI+iFgpXIiIiIiKlq1DhqlWrVrz//vusW7eOZcuWcd111wFw7NgxqlevXqwNlCuoFgI2V0g8BglZ8+AUrkRERERESlehwtWrr77KBx98QJ8+fbjtttto164dAIsWLcocLiilxM0HqrY1+xf1XilciYiIiIiULrfCvKhPnz6cOnWKuLg4/P39M4/ff//9+Pj4FFvjJJ9qdIOz2024qneLOVTDnDp6FJKSwMvLge0TEREREakACtVzlZiYSHJycmawOnz4MDNmzGDv3r0EBAQUawMlH3IpatGsGQQFwfnzsGiRg9olIiIiIlKBFCpcDR8+nLlz5wIQExND165deeONNxgxYgSzZs0q1gZKPmQUtTizFeypALi6wl13mcOzZzuoXSIiIiIiFUihwtW2bdvo1asXAN9++y2BgYEcPnyYuXPn8vbbbxdrAyUfqjQFD39IT4KYnZmHx441j0uXmuGBIiIiIiJScgoVrhISEqhSpQoAS5cu5aabbsLFxYVu3bpx+PDhYm2g5IPNBtVzLibcuDH06gV2O3z2mYPaJiIiIiJSQRQqXDVp0oSFCxcSGRnJkiVLGDhwIAAnTpzA19e3WBso+ZTHeld3320eZ88GyyrlNomIiIiIVCCFClfPPfccTz31FA0aNKBLly50794dML1YHTp0KNYGSj5lFrXYlO3wP/4BlSrBX3/Bxo0OaJeIiIiISAVRqHB1yy23EBERwZYtW1iyZEnm8X79+jF9+vRia5wUQI0L64ud2wfJpzMPV64Mt5jq7MyZU/rNEhERERGpKAoVrgBq1apFhw4dOHbsGEcvVEvo0qULwcHBxdY4KQAPf/BtbvZPZe+9yhga+NVXkJBQyu0SEREREakgChWu7HY7U6dOxc/Pj/r161OvXj2qVq3KSy+9hN1uL+42Sn7lst4VmKIWjRrBuXOwYIED2iUiIiIiUgEUKlxNnjyZd999l1deeYXt27ezbds2/vvf//LOO+8wZcqU4m6j5FdmUYvsPVcuLlrzSkRERESkpNksq+A15OrUqcP777/PDTfckO34999/z8MPP5w5TNBZxcXF4efnR2xsrHNVPzyzHX7pCO5+cMsZsGVl58OHoWFDUzHw0CGoX99xzRQRERERcRYFyQaF6rk6c+ZMrnOrgoODOXPmTGFuKcWhahtw9YbUWIjbm+1U/fpw7bVm/9NPHdA2EREREZFyrlDhql27drz77rs5jr/77ru0bdu2yI2SQnJxg+qdzf4l610BjB1rHufMMQsLi4iIiIhI8XErzItee+01hg4dyvLly+nevTs2m40NGzYQGRnJ4sWLi7uNUhDVu8GJtWa9q8Z3Zzt1003wyCNw8CCsWwfXXOOgNoqIiIiIlEOF6rm65ppr+Ouvv7jxxhuJiYnhzJkz3HTTTfz555/MVsUEx6rR1Tzm0nPl4wOjRpl9/WMSERERESlehSpokZcdO3bQsWNH0tPTi+uWDuG0BS0AEo7BwqtMMYtbYsG9crbTGzdCjx4maEVHQ5UqDmqniIiIiIgTKPGCFlKG+dQBnyCw7HBmS47T3bpB8+ZmMeFvv3VA+0REREREyimFq/IoY72r05tynLLZsgpbaGigiIiIiEjxUbgqj6pnLCacc94VwOjRZmHhdetg//5SbJeIiIiISDlWoGqBN91002XPx8TEFKUtUlwuLmphWaa76iJXXQUDB8Ivv5g1r156yQFtFBEREREpZwrUc+Xn53fZrX79+owZM6ak2ir55d8RbG6QFA0JEblecveFKu2ffgpOXn9ERERERKRMKFDPlcqsOwk3b/BvbwpanNoElernuOSGG8DfHyIjYeVKGDCg9JspIiIiIlKeOHzO1cyZM2nYsCFeXl6EhISwbt26PK+Niori9ttvp3nz5ri4uDB+/Pgc18yZMwebzZZjS0pKKsFPUQbVuPy8Ky8vuO02sz9nTuk0SURERESkPHNouAoNDWX8+PFMnjyZ7du306tXLwYPHkxERO5D2ZKTk6lZsyaTJ0+mXbt2ed7X19eXqKiobJuXl1dJfYyyqXrGvKuNeV6SMTRwwQLQdDkRERERkaJxaLh68803GTduHPfeey8tWrRgxowZBAUFMWvWrFyvb9CgAW+99RZjxozBz88vz/vabDZq1aqVbatwal5tHk9vgpg/c70kJARat4akJAgNLcW2iYiIiIiUQw4LVykpKWzdupWBAwdmOz5w4EA2bNhQpHvHx8dTv3596taty/XXX8/27dsve31ycjJxcXHZNqdXuQEE3QRY8McLuV6iNa9ERERERIqPw8LVqVOnSE9PJzAwMNvxwMBAoqOjC33f4OBg5syZw6JFi5g3bx5eXl707NmTffv25fmaadOmZat6GBQUVOj3L1PavAjYIPJbOJN7wLzzTnB1hU2bIDy8dJsnIiIiIlKeOLyghe2SNZgsy8pxrCC6devGnXfeSbt27ejVqxdff/01zZo145133snzNZMmTSI2NjZzi4yMLPT7lylVW0P9C1Urdj6X6yWBgTB0qNlXYQsRERERkcJzWLiqUaMGrq6uOXqpTpw4kaM3qyhcXFzo3LnzZXuuPD098fX1zbaVG22eB5sLHPvRlGXPRcbQwLlzIS2t9JomIiIiIlKeOCxceXh4EBISwrJly7IdX7ZsGT169Ci297Esi7CwMGrXrl1s93Qqvs2g4V1mf+eUXC8ZOhRq1IDoaFi6tBTbJiIiIiJSjjh0WOCTTz7J//3f//HJJ58QHh7OhAkTiIiI4MEHHwTMcL0xY8Zke01YWBhhYWHEx8dz8uRJwsLC2L17d+b5F198kSVLlnDgwAHCwsIYN24cYWFhmfeskFo/By7uEL0MTqzNcdrDw8y9AhW2EBEREREpLDdHvvmoUaM4ffo0U6dOJSoqitatW7N48WLq168PmEWDL13zqkOHDpn7W7du5csvv6R+/focOnQIgJiYGO6//36io6Px8/OjQ4cOrF27li5dupTa5ypzKjeARuNg//uw41nov8aUCrzI2LEwYwYsWgSnT0P16o5oqIiIiIiI87JZlmU5uhFlTVxcHH5+fsTGxpaf+VcJR2BRE7AnQ9+lUHtAjks6doTt2+Gdd+DRRx3QRhERERGRMqYg2cDh1QKllPjUhaYXhkbunAK5ZOq77zaPGhooIiIiIlJwClcVSctnwNUHTm+CYz/lOH3bbeDuDtu2wc6dDmifiIiIiIgTU7iqSLxrQfPHzP7OKWDZs52uUQNuuMHsa80rEREREZGCUbiqaFr8C9yqwNkwiPwux+mMoYGffw6pqaXbNBERERERZ6ZwVdF4VofgCWb/j+fAnp7t9KBBUKsWnDwJ33/vgPaJiIiIiDgphauKKHgCePhD7G6ICM12ys0N7r3X7L/9tgPaJiIiIiLipBSuKiKPqtDiKbO/83mwp2U7/dBDJmStW2eKW4iIiIiIyJUpXFVUzR4HzxoQvx8Ozs12qk4dGDXK7L/1lgPaJiIiIiLihBSuKir3yqY0O8CuqZCeku30E0+Yx3nzIDq6lNsmIiIiIuKEFK4qsqYPgXdtOH8YDnyc7VTnztC9u6kY+P77DmqfiIiIiIgTUbiqyNx8oNVks7/rZUhLzHZ6/HjzOGsWJCeXbtNERERERJyNwlVF1/he8AmCxGOw/4Nsp268EerWhRMn4KuvHNQ+EREREREnoXBV0bl6QuvnzP7uaZAan3nK3R0efdTsv/UWWJYD2iciIiIi4iQUrgQa3QWVG0HSCfjr3Wyn7rsPvL1h+3ZTml1ERERERHKncCXg4g5tXjD74a9BSmzmqWrVYPRos6+y7CIiIiIieVO4EqP+7eAbDClnYe+MbKcyyrIvXAiHDpV2w0REREREnIPClRgurtDmRbO/501IPpN5qmVLGDAA7HZ49908Xi8iIiIiUsEpXEmWerdA1baQGgfh/8t2KqP36v/+D+Ljc3mtiIiIiEgFp3AlWWwu0PYls7/3LVPg4oLBg6FpU4iNhblzHdQ+EREREZEyTOFKsrtqGPh3hPQEOPRl5mEXF3j8cbP/1ltmiKCIiIiIiGRRuJLsbDZTmh0g8ttsp8aOBT8/+OsvWLKk9JsmIiIiIlKWKVxJTkE3mceT6yHhaObhypVh3DizP2NG6TdLRERERKQsU7iSnHzqQo0eZj9yfrZTjz5qhgguXQrh4Q5om4iIiIhIGaVwJbmr9w/zGPFNtsMNG8Lw4Wb/7bdLuU0iIiIiImWYwpXkrt4t5vHkekg4lu1URln2Tz+FM2cQEREREREUriQvPnWhRnfAgsgF2U717g3t20Nioln3SkREREREFK7kcjKGBkZmHxpos2X1Xr37LqSllXK7RERERETKIIUryVvQzebxxDpIjMp26tZboWZNiIyE775zQNtERERERMoYhSvJW6V6UL0ruQ0N9PKChx4y+2+9VfpNExEREREpaxSu5PLyqBoIJly5u8P69bBlSym3S0RERESkjFG4ksvLqBp4Yi0kHs92qlYtMzwQ1HslIiIiIqJwJZdXqT5U7wJYcGRBjtMZhS1CQyEqKsdpEREREZEKQ+FKruwyQwNDQqBnT0hNhVmzSrldIiIiIiJliMKVXFlm1cA1kHQix+nx483j++9DUlLpNUtEREREpCxRuJIrq9wQqnUCy56jaiDAiBFQrx6cPAnz5pV+80REREREygKFK8mfzKGB3+Y45eYGjzxi9t96CyyrFNslIiIiIlJGKFxJ/mSEqxOrIOlkjtP33gs+PrBjB6xdW8ptExEREREpAxSuJH8qN4RqIWZo4JHvcpyuVg3GjDH7M2eWcttERERERMoAhSvJv8tUDQS45x7zuHixCluIiIiISMWjcCX5F3RhQeHjqyDpVI7TISFw1VUQHw8rV5Zy20REREREHEzhSvKvSmPw7wBWOhxZmOO0iwsMH272v/++dJsmIiIiIuJoCldSMFcYGnhxuLLbS6lNIiIiIiJlgMKVFExGuDq+ApJP5zjdpw/4+sLx47BpU+k2TURERETEkRSupGCqNAH/9nkODfTwgKFDzf7CnKdFRERERMothSspuIzCFvkYGigiIiIiUlEoXEnBZQwNjF4ByWdynB48GNzdYe9e2LOnlNsmIiIiIuIgCldScL7NoGpbsNLgSM7uKV9f6NfP7GtooIiIiIhUFApXUjgFqBooIiIiIlIRKFxJ4WRWDVwOKWdznL7hBvP4228QFVWK7RIRERERcRCFKykc3+ZQtQ3YU3MdGlinDnTtavYXLSrltomIiIiIOIDClRSeqgaKiIiIiGRSuJLCy6wauAxSYnKcHjHCPK5YAXFxpdYqERERERGHULiSwvNrAX6tLgwNzDn2LzgYmjWDlBT45RcHtE9EREREpBQpXEnRXKZqoM2moYEiIiIiUnEoXEnRZA4NXAopsTlOZwwN/Okn04MlIiIiIlJeKVxJ0fi1NJs9BY7mHBrYtSsEBkJsLKxZ44D2iYiIiIiUEoUrKbrMqoHf5jjl6grDhpl9DQ0UERERkfJM4UqKLmNoYNQSSM1ZFjBjaODChWBZpdYqEREREZFSpXAlRefXCnyDwZ4MR37IcbpfP6hUCY4eha1bHdA+EREREZFSoHAlRWezZfVeReasGujlBdddZ/Y1NFBEREREyiuFKykeGeHq2C9XHBooIiIiIlIeKVxJ8fBrDb7NzdDAoz/lOD10qClusWsX/P23A9onIiIiIlLCFK6keNhsWVUDD3+V47S/P1xzjdnX0EARERERKY8UrqT4NLjDPB5bDInHc5zW0EARERERKc8UrqT4+LWA6l3BSoNDn+c4PXy4eVy/Hk6eLOW2iYiIiIiUMIUrKV6N7jaPB2bnWNSqXj3o2BHsdvjxRwe0TURERESkBDk8XM2cOZOGDRvi5eVFSEgI69aty/PaqKgobr/9dpo3b46Liwvjx4/P9br58+fTsmVLPD09admyJd99910JtV5yqH8ruHpB7J9wZkuO0xm9VxoaKCIiIiLljUPDVWhoKOPHj2fy5Mls376dXr16MXjwYCIiInK9Pjk5mZo1azJ58mTatWuX6zUbN25k1KhRjB49mh07djB69GhGjhzJpk2bSvKjSAYPP6h7k9k/MDvH6Yx5V0uXwvnzpdcsEREREZGSZrOsS8ZulaKuXbvSsWNHZs2alXmsRYsWjBgxgmnTpl32tX369KF9+/bMmDEj2/FRo0YRFxfHzz//nHnsuuuuw9/fn3nz5uWrXXFxcfj5+REbG4uvr2/+P5AY0cth5QBwrwo3RZmerAssCxo3hoMH4bvvssKWiIiIiEhZVJBs4LCeq5SUFLZu3crAgQOzHR84cCAbNmwo9H03btyY456DBg267D2Tk5OJi4vLtkkRBF4LPvUgNQYiF2Y7ZbNpaKCIiIiIlE8OC1enTp0iPT2dwMDAbMcDAwOJjo4u9H2jo6MLfM9p06bh5+eXuQUFBRX6/QWwuUCjsWb/MkMDf/gB0tJKrVUiIiIiIiXK4QUtbDZbtueWZeU4VtL3nDRpErGxsZlbZGRkkd5fyApX0cvgfPafZ8+eUL06nDkDv/5a+k0TERERESkJDgtXNWrUwNXVNUeP0okTJ3L0PBVErVq1CnxPT09PfH19s21SRJUbQkAfwIKDc7OdcnOD6683+99/X+otcxzLgviDYNkd3RIRERERKQEOC1ceHh6EhISwbNmybMeXLVtGjx49Cn3f7t2757jn0qVLi3RPKaTMNa/m5FjzKmNo4MKFOU6VXwfmwKJGsPdtR7dEREREREqAQ4cFPvnkk/zf//0fn3zyCeHh4UyYMIGIiAgefPBBwAzXGzNmTLbXhIWFERYWRnx8PCdPniQsLIzdu3dnnn/iiSdYunQpr776Knv27OHVV19l+fLlea6JJSWo3s3gVgXi98PJ7OP/Bg4Eb284dAh27nRM80rdoS/MY+QCx7ZDREREREqEmyPffNSoUZw+fZqpU6cSFRVF69atWbx4MfXr1wfMosGXrnnVoUOHzP2tW7fy5ZdfUr9+fQ4dOgRAjx49+Oqrr3j22WeZMmUKjRs3JjQ0lK5du5ba55IL3CpB/ZHw98dw4BMI6JV5yscHBgyARYvM0MA8li0rP9IS4OSFBbLPbAZ7Kri4O7ZNIiIiIlKsHLrOVVmlda6K0cn1sOxqE7RujAb3ypmnZs+Ge+6BDh1g2zYHtrE0HFsCq6/Lej7od6je2XHtEREREZF8cYp1rqSCqNEDqjSDtPMQ8U22U9dfDy4usH07HD7soPaVlujs8wA5Wfi13ERERESkbFK4kpJls+W55lXNmqYsO5jhgeVa9FLzWLWNeTy10XFtEREREZESoXAlJa/hGLOw8Ml1cG5/tlMXVw0stxKjIeYPs9/qWfN4Sj1XIiIiIuWNwpWUPJ+roNZAs39gTrZTw4ebxzVrzKLC5VL0cvPo3xHqDDFBMyESEo44tl0iIiIiUqwUrqR0ZKx5dfBTsKdnHm7cGFq3hvR0WLzYQW0raRnzrWoPMAU9ql4ojaihgSIiIiLlisKVlI66N4CHv+mtOb4i26lyPTTQsrLCVa0B5rHGhQWtVdRCREREpFxRuJLS4eoF9W83+5cUtsgIVz/9VA4DVuyfkBgFrt5Q80L1jhrdzaPmXYmIiIiUKwpXUnoa32MeI7+DlLOZhzt2hGuvhaQkuPFGeOQRSEx0UBuLW0avVUBvEzABal7ouTq7HdLKywcVEREREYUrKT3+HaBqW7Anw+GvMg/bbPDzz/Cvf5nnM2dCly7w558OamdxirpQgj1jSCBApQbgVQvsqXBmq0OaJSIiIiLFT+FKSo/NllXY4u/sQwM9POC112DJEggMhF27oHNn+PBDM23JKaUnw4k1Zv/icGWzaWigiIiISDmkcCWlq8EdYHODM5shZleO0wMHwo4dMGiQGRr4wAPwj3/A2bO53KusO7UB0hPBKzBr8eAMGUMDVTFQREREpNxQuJLS5VUTrhpm9i8pbJEhMNCUZf/f/8DdHebPh3bt4NdfS7GdxSHqoiqBNlv2cxkVA09tcOKuORERERG5mMKVlL6MoYGHPjfzjnLh4gL//Cds2ABNmkBkJFxzDUydatbEcgrRucy3ylCtI7i4Q9IJiD9Quu0SERERkRKhcCWlr85gM1Qu6QQcu/zKwZ06wbZtMGYM2O3w/POmsuCRI6XU1sJKPg1ntpn92rmEK1cv8A8x+xoaKCIiIlIuKFxJ6XNxg4ajzX4eQwMvVqUKfPopfPYZVK4Ma9eaYYJlek2s6BWABX6twbt27tfUvGhooIiIiIg4PYUrcYyMoYFHfzI9WPlw552wfbvpzTpzpoyviXW5IYEZMuZdnVS4EhERESkPFK7EMfxaQvUuYKXBwc/z/bImTWD9enjqKfM8Y02sA2Vp2pJlZRWzqD0w7+syyrHH/gGp50q+XSIiIiJSohSuxHEyeq8OzC5QxTwPD3j9dfjlFwgIMGti9eoF4eEl1M6COrcPEiLAxQMCeud9nU8dqFQfLDuc/r302iciIiIiJULhShyn/q2msEPsLjiztcAvHzQIwsKgdWs4dgx69zbPHS76Qq9VzZ7g5nP5azU0UERERKTcULgSx/GoCnVvNPv5KGyRm9q1YfVqCAmBU6egb1/47bdia2HhROVjvlWGjKGBqhgoIiIi4vQUrsSxMte8+hLSkwp1i+rVYcUK6NEDYmJgwAATuBzCngrHV5n9y823ypBZMXCjGR4oIiIiIk5L4UocK/Ba8KkHqTEQubDQt/Hzg6VLoV8/iI+HwYPh55+LrZX5d/p3SDsHntXBv8OVr6/aFlx9zOeP21PizRMRERGRkqNwJY7l4gqN7jL7Bz4p0q0qVYIff4Trr4ekJBg+HBYsKIY2FkTGkMDAfmDLx79eLu5QvbPZ19BAEREREaemcCWO12iseYxeBhvGFKksuZeXCVQjR0Jqqnn8PP+V3osuOh8l2C+lohYiIiIi5YLClThe5UbQ4XXT03PoM/i5A5wqfGlyd3f48ksYOxbS02HMGPjww+Jrbp5SYrNKquenmEWGzHlXClciIiIizkzhSsqGFk9BvzVm/lX837CsJ/z5SqGLPLi6wscfwyOPmCW0HngApk8v5jZf6vgqsNKhSjOoVC//r6vezTzG7YHkMyXTNhEREREpcQpXUnYEXA1DwqDeP8BKgx2TYOUASDhaqNu5uMA778DTT5vnTz4JL71UoPWKCyb6wnyrggwJBPCqYQIZwClH15EXERERkcJSuJKyxcMfeoZC149NFb3jK2FxWzjyfaFuZ7PBK6/A1Knm+XPPwaRJJRSwoi7MtyrIkMAMGhooIiIi4vQUrqTssdmg8T0weBv4d4SUM7B2BGx+GNISC3W7KVPgjTfM81dfhccfB3txLisVfxDi94PNFQL7FPz1WkxYRERExOkpXEnZ5dscBm4w87EA9s2CJZ0g5o9C3e7JJ+GDD0zYevddGDfOFLwoFhlVAmt0A3ffgr8+o2Lg6U1gTyumRomIiIhIaVK4krLN1dNUEuy7BLwCIXY3/NIZ9r5bqLF9998Pc+ea+Vhz5sDAgbBjRzG0M3NIYAHnW2Xwa2lCWdr5QodHEREREXEshStxDrUHwpCdUGco2JNh62Ow5gZIOlngW915J3zzDXh4wMqV0KGDOXbwYCHbZk+H4yvMfmHmW4EpQ59RNVBDA0VEREScksKVOA+vALjmBwh5G1w84diPpthFRq9RAdx0E/z5J9x6q+kA++ILaN4cnngCThY0r53dBilnwd0PqncucFsyqaiFiIiIiFNTuBLnYrNB88dg0O/g2wKSomHVQNh0b4F7sZo0gXnzYMsWGDAAUlPh7behUSNTXTA+Pp83irpQgj3wWnBxK9jnuVjGvKuTClciIiIizkjhSpyTf1u4bgs0fcg8//tj+KGZmYtVwIIQISGwdCksW2b24+Ph+eehcWN47z1ISbnCDTKKWdQu5JDADDW6AjY4fxASo4t2LxEREREpdQpX4rzcfKDzTBjwK/i3h9QYMxfrlxA4sa7At+vfH37/HUJDTa/WiRPw6KPQogV89VUepdtT47OG8RV2vlUGd1+o2trsa96ViIiIiNNRuBLnV7MnDNpigpaHP8TshOW9YcOdkHCsQLdycYGRI2H3bpg5EwID4cABuO026NTJ9G5lc2IN2FOhUkOo3Ljon6WG5l2JiIiIOCuFKykfXFzNEMHr/4ImDwA2OPQF/Ngcdr8O6Vca25eduzs89BDs3w8vvQRVqsD27aZ0e//+Zp4WkH1IoM1W9M+hxYRFREREnJbClZQvXjWgy/tw3WZT2jwtHsKehp/bZhWeKIDKleHZZ03v1fjxpnz7ihXQuTOMHQupkRnrWxVxSGCGzMWEt0B6cvHcU0RERERKhcKVlE/VQmDgeug225Rwj9sLqwbBupvh/OEC365GDZg+HfbuNWtiASxbdBT3hN1Ylo2kqtcWT7urNAHPGmYtr7Pbi+eeIiIiIlIqFK6k/LK5QKOxZqhg8/Fgc4XIBfBjMPwxFdISC3zLBg3gs89g0yZ4cLjptfr978606lCNhQvNmllFa7Mta2igSrLLxRKPQ+TCYvglExERkZKicCXln4cfhEyHwWEQ2BfSk+CP5+GnVhAxv1BfVrt0gWfvNeHqt0MDOHAAbrzRrJe1a1cR25tZ1ELzruQi62+FdTfCwU8d3RIRERHJg8KVVBxVW8O1K6BnKPjUNetJ/XoL/NIJjv1csJBl2bEdXw7Afc8NYPJk8PQ087HatTMl3E+fLmQ7a15UMVC9FAIQ8wecWG3293/o0KaIiIhI3hSupGKx2aD+SBgaDq2fA7fKcHYbrB4Cy3vB8dX5u0/MTkg6AW6V8AnqzssvQ3g43HyzWQ/rvfegaVN45x1ITS1gG6t1ApsbJB6DhIiCfkIpj/a9n7V/aiPEhjuuLSIiIpInhSupmNwrQ9sX4YYD0OIpcPWCk+thRV9YOQBObbr866MuVAkM6AOuHgA0bAjffgsrV0LbtnD2LDz+OLRvD8uXF6Btbj5mUWSAkxoaWOGlxsPBz8x+pfrm8cAnjmuPiIiI5EnhSio2r5rQ4XUY9jc0fRhc3CF6OSztBmtugLM7cn9ddN4l2Pv2ha1bYdYsqF7dLEg8YACMGGHWzcoXLSYsGQ59AWnnoEoz6DjDHDs41yxeLSIiImWKzbI0qeNScXFx+Pn5ERsbi6+vr6ObI6Up/hDsmmqKBlh2c6zeSGjzIvgFm+dpifCtvymXPnQ3+LXI83Znz8LUqfDuu5CWBm5ucNVV4OcHVauaLWP/4mOtKn9Ft/TbOO8ZwrG2W6ha1QQ1F/05pGKxLPi5A8TsgI5vQrNHYWEQJB2HXt9B0AhHt1BERKTcK0g2ULjKhcKVELcXdj4PEaHmuc0FGo4x87TO7YdVA8H7KhgRaeZxXUF4OEyYAEuW5O/tg6pHEPF2fdLSXfG7L5aE5Eo0aABPPQX33APe3oX/aOJETm6EZT3MsNURR8GzGmx/GsJfh6uGwTWLHN1CERGRck/hqogUriTT2Z2wcwocvfAl1sUdKjeGuD1mDa1us/N9K8uCw4fhxAmIjYWYGLPlvm/x7R1BBPoeZfAbq/ll2zWZ96lZE8aPh4cfNr1cUo5tGAOHPsv+uxa7B35qYdZtGxEJ3rUd2kQREZHyTuGqiBSuJIdTm0zIyphrBdDjC2hwe8m957p/QOS30O6/JDScxJw58PrrcOiQOV2lCjz0kAlatfX9uvxJPg3fXWWGnw7cBDW6ZJ1b2tPMx2v/CrSc6Lg2ioiIVAAFyQaawSGSHzW6wrVLod9qCLwWanSHq64v2fesmbWYsI+P6anatw8+/xxat4Zz5+C116BBA3jggQIUyxDncGCOCVb+HaF65+znGt9jHv/+RGuhiYiIlCEKVyIFEXgN9FsBAzeAewn3atbIuZiwmxvccQfs3Ak//AA9ekBKCnz4ITRvDrfeCtu3l2yzpBRY9qy1rZo+lHNeX72R4FYJzv1llhAQERGRMkHhSqSs8u8ALp5meNi5fdlO2Wxw/fWwfj2sWwdDh5rFi0NDoWNHuO46WLNGnRpOK3oFxO83Ab7BbTnPu1cxAQu05pWIiEgZ4uboBohIHlw9oHon0zNxaiP4Nsv1squvhh9/NL1Zr74KX31lqhIuWQLdupnhhL6+JnxlbOnp2Z/ndtzFxYS0+vVL+XML7JtlHhuOMT1UuWk8Dg7MhoivIeQtE7hERETEoRSuRMqyGj0uhKsN0Oiuy17ati188QW89BK88QZ8/DFs2mRn0yawrMJ1Unt7m3W6xo83QxKlFCQczapO2eTBvK+r0QN8m5tlAyK+NmFLREREHEpfl0TKshrdzePJDfl+SaMG6bz3/HpeH/UtqQcX4GU7zRfb/sXcrZNJxxNXV9MrdfGW27GICNiyBf71LxPaPvoIOnUqoc8pWfZ/BFY6BPSGqq3yvs5mg0b3QNhE+PtjhSsREZEyQKXYc6FS7FJmJEbDd7UBG9xyFjz8cr/OngrHV0PkfDjyHSSdyHmNX0vo+ompfJgPlgWzZ5uFi8+eNYHr8cdNz1jlyoX+RHI59lT4vgEkHoMe86DBrZe/PjEKFgaZMDZ0N/i1KJVmioiIVCQqxS5SXnjXgsqNAAtOb8p+Lj0Zjv4Ev90DC2rBqoGw/wMTrNyrQsO7oPci6PkVeAVA7G5Y1gO2/RPSEq741jYb3HMP7NkDt99u5mHNmAGtWpk5XlICjv5ggpVXAATddOXrvWtDnSFmX4UtREREHE7hSqSsu3hoYFoiRH4HG+6EBQGw5npT1CDlDHjWhMb3Qd8lcPMJ6D4H6g6D+qNMr0aDO02J7z1vwuK2cHxNvt4+IMAMC/z5Z7OmVkQEDBsGI0dCVFSJfeqKKaP8eqNxpqBJfmQMBzw41/R8iYiIiMNoWGAuNCxQypS/ZsKWR8CzhulxSr+o18m7NtS9CerdDDV7gcsVplEe/Ql+fwASj5rnTR+C9q/ke82u8+fhxRfhzTdNZUE/P7OQ8b33mmGDUgRx++DHZoANbjgAlRvk73X2VDM0MOk49F4IdYeXYCNFREQqHg0LFClPavY0j8mnTLDyqQfNJ8CA9TDiCHR+FwL7XjlYAVw1FIb+CU3uN8/3zYKfWsOxX/LVlEqVTJjavBlCQiA2Fh54AHr3ht27C/n5xNj/gXmsMyT/wQrAxd2UbAdT2EJEREQcRj1XuVDPlZQ54W9CymmoeyNUCzEToooqeiVsuhfOHzTPG46BjtPBs1q+Xp6eDu++C5Mnmx4td3eYNMlsXl5Fb16FkpYIC+ua4Z3X/ABXXV+w18fugZ9agM0VRkSaHk0REREpFuq5EilvWjwJ7f5jFhUujmAFUOtaGPoHNB8P2MycnZ9aQuSCfL3c1RWeeML0WA0dCqmpZk2s9u1h0SLTqyX5FPGNCVaV6kPtwQV/vV+wWffKSjf/HEVERMQhFK5EKjK3ShAy3Qwx9G1h5u2suxnW/QMSj+frFvXqwQ8/wNdfQ61asHcvDB8O/v7QsiXcfTe8/z5s324CmORi3yzz2OR+cHEt3D0a32Me//7E1NEXERGRUqdhgbnQsECpkNKTYNdLsPtV0wPiUc30atW72QSvfPSYxcTACy/A99/DoUM5z3t7Q8eO0LVr1lavXvF1xjmls2Hwcwczd2p4JHgHFu4+qefMmmhp56H/Ogi4ulibKSIiUlEVJBsoXOVC4UoqtDPbYdM95kt/Bt/mEHSzWXvJv2O+0tDx4/D772bbtMk85jZUMDDQhKwuXczWrBnUrWuGHVYIvz8A+z+EeqPg6q+Kdq/f7jGl+RvdDd207pWIiEhxcKpwNXPmTF5//XWioqJo1aoVM2bMoFevXnlev2bNGp588kn+/PNP6tSpw9NPP82DDz6YeX7OnDncfffdOV6XmJiIVz5n2StcSYVnT4VDX0DEtxC9DOwpWed86pmQFXSzWYMrn8PY7Hb46y8TtDK2nTshLS3nte7uZk2tRo2gcWPzePF+5crF8zEdLjUOvqtjepv6rYbAa4p2v5PrYdnVZrjnjVHgXqU4WikiIlKhFSQb5KN2c8kJDQ1l/PjxzJw5k549e/LBBx8wePBgdu/eTb169XJcf/DgQYYMGcJ9993H559/zvr163n44YepWbMmN998c+Z1vr6+7N27N9tr8xusRAQzRK3RWLOlxpn1sSIXwLHFkBABe2eYzSsQ6o4wYSuwr3ldXrd0geBgs911lzmWmGjmYmWEre3b4eBBMzdr3z6z5SYgIGfgatUK2rRxskqFBz83wcqvJQT0Lvr9avQwvYxxeyHi66wFhkVERKRUOLTnqmvXrnTs2JFZs2ZlHmvRogUjRoxg2rRpOa6fOHEiixYtIjw8PPPYgw8+yI4dO9i4cSNgeq7Gjx9PTExMvtuRnJxMcnJy5vO4uDiCgoLUcyVyqbREiF4KEfPh6A+QGpN1zr0q1L3BBK1aA8HNu1BvkZ4OR4/C33/DgQNZjxn7Z85kXWuz2bmp8wIeHfAuLi529hxryVl7S1yrtaB6w5Y0b1+Hdu1tVKpUpE9dMiwLFreF2F0Q8jY0f6x47rv7NQibaILWwPXFc08REZEKzCl6rlJSUti6dSvPPPNMtuMDBw5kw4YNub5m48aNDBw4MNuxQYMG8fHHH5Oamoq7u/mreXx8PPXr1yc9PZ327dvz0ksv0aFDhzzbMm3aNF588cUifiKRCsDNG+oON1t6CpxYbXq0jnwHSSdMGfCDc03QavoANHsUfOoW6C1cXU2Ri3r1oG/fnOdjYuDA33aS/15Ak8QXqem+K/Nc7+B12a6NDffljxUtOHa+JcmeLfAObEnt4Ja06FQfXz8HF0s9ud4EK1cfaDi6+O7bcDTs+Dec2gCx4eDXovjuLSIiIpflsHB16tQp0tPTCQzMXhkrMDCQ6OjoXF8THR2d6/VpaWmcOnWK2rVrExwczJw5c2jTpg1xcXG89dZb9OzZkx07dtC0adNc7ztp0iSefPLJzOcZPVcichmuHlB7oNk6vWe+zEcugMj5kBBpqg6GvwH1b4XgCVCtY9Hf07JTNW4+HaOnQtoucAfcfaH5eKwqzTh3ZDfxx8JxT9hNNff9+PnE0a3JJmBT1j2OQ8ICb/48E8zZ9BYke7flnPfVWFU7Ua2mJzVqQI0aUL06uJXkfyEzyq83uA08qhbffb1rQ50hpmfxwGzo8Frx3VtEpAAsy8yrdc97xLhIuePQOVcAtkuqjlmWlePYla6/+Hi3bt3o1q1b5vmePXvSsWNH3nnnHd5+++1c7+np6Ymnp2eh2i8imKIWAb3M1vENOPoj7HkTTqyBQ5+bLaAPBD8JVw0FWwF7jSy7CW1/vAixf5pj7r7QfAIEPwEe/tgA34aQ2Vmfngzn9nP28G5O7t9N6qlwKqXvpnblvfh4JtKq9nZgO/AlAInxXvwW1o35e3qxdk9vNu7vjod3pcywVaMG1KyZtR8QAH36mMIbBZZ0EiK/NftNHyrEDa6g8TgTrg5+ahafvsxcuBJhWXDuLzP/S0QqJMsy82u/+QamTIGnny7hP1iJlBEO+zWvUaMGrq6uOXqpTpw4kaN3KkOtWrVyvd7NzY3q1avn+hoXFxc6d+7MvrxmxotI8bK5mLlXdW+AM1thz3Q4HGqGEJ5YDVWamp6shmNMVbvLyTVU+Zn1ty6Eqjy5ekLVVvhXbYV/u4uO29M4E3GQw3/sJjZiN1VSt9DEdy1+Xqfo23I1fVuuBiA1zY1thzqydk9v1u7pzfpVPTl7vlqOt+ncGf7xD7PlO2gd+MRUYKzWGaqFXPFyyzLFPQ4eNIU7rrrqCtXw6wwxxUaSjpsiJHWH57NhxcCeDr/+wwwVbfVvE+5EpMJZsAA++8zsT54MP/4In34KeQwiEik3HF7QIiQkhJkzZ2Yea9myJcOHD8+zoMUPP/zA7t27M4899NBDhIWFZRa0uJRlWXTp0oU2bdrwySf5W/dFpdhFilnCEfjrXdj3QVYRDI9q0PRBaPoI+NTJfn1eoSp4AjR/oniH0YFJL3F74eRaOLEW68RabAmROS47kdKGfTG92RHVi6VhvfhheW3s9qyU06lTVtBq2DCv97LDoiZw/iB0/QQa51w6AiAyElauzNqOHMk6FxAAISHZt7p1Lwlc25+G8NfhqmFwzaJC/FAKacsT8FfGKAEb9FsJgX1K7/1FSlNiFEQvN4V8rvTHogokJgZatoSoKBg6FNatg7g48PGB//0PHnywgi8eL07Hada5Cg0NZfTo0bz//vt0796dDz/8kI8++og///yT+vXrM2nSJI4ePcrcuXMBU4q9devWPPDAA9x3331s3LiRBx98kHnz5mWWYn/xxRfp1q0bTZs2JS4ujrfffpvPPvuM9evX06VLl3y1S+FKpISkxsOBObB3OsQfMMdc3KH+bSY4VW1buqHqcs4fhhNr4cQ6E7ri9ua4xO5amZiU+uw7Vo/tf9Xn0Mn6RJyux+FT9alaux7XDK7DLf9wpVGji1507BdYPdgU/bjxKLj5AHDyJKxalRWmLu1s9/AwPWN//20qKl6qZs3sYatriz3U2dYCbK4wItLMxSppe96CbePNfvVucPo38AmCITtL9J/d6dMwbRrs32++uDVpUmJvJfmVcAyWX2OWGei1IN/r4TmV+IPmMyZEgm8w9AwF/7aOblWZ8NBD8P77ZlH4HTvMovJ3323+GwcwaBB8/LHphRdxBk4TrsAsIvzaa68RFRVF69atmT59Or17m/Vexo4dy6FDh1i9enXm9WvWrGHChAmZiwhPnDgx2yLCEyZMYMGCBURHR+Pn50eHDh144YUX6N69e77bpHAlUsLs6XB0kZmXdfLXrONetSDpwtBfR4WqvCQeN209sdaErbM7gMv/5zM1zY0jZ+pyJrkentXqc1XTevgnr4DTv5HccDzLTk9nxQoTpnbuzP5aFxcz5PDaa83Wo4f5q29iovmysnVr1vbnn7kHrk0v9aRLow38cvwVdlsTcXc3cx7c3cm2n9eju7vpEatTJ+e9c4j8DtbdbH4m7V+Fpg/Dzx0gfr8Jzz2/zO9POt8SE+Htt02wio01x/z84IsvzF/LxYF+vRUiQs1++1eg5UTHtqe4nT9sgtX5w1nHXDyh45tmHmUF7pb59Vfo1cvsr14N11xYG91uh3fegWeegaQkqFoVZs6E225zVEtF8s+pwlVZpHAlUopO/W56siK+ASu97IWqvKQlmr9Ynz9stoSIC/sRpJ87DAmRuNrS8nx5y6f3EH40e8GHNm2gXz8Tpnr3NkEhPxITTTi7NHCN6fkxH99/L3uPNSP4X3uAgn/hs9nguuvMMJ4hQ/KYkH5qE6zoA+lJ0ORB6DzTvPDUJljW0/xz7fEFNLi9wO+fG7sdPv8cnn3WDJ8EaN8ePD3NYtQAL7xgJtG7lGTF/YQjgEvOYa0VXdRSWDUo67mLOwz8rXgqhpYF5yNgeR8ztLdKU+g1H8L+Dcd+NOfr3ghd/w88c87RLO+Sk6FDBwgPh3Hj4P/+L+c14eEwZgxs2WKejxoF771nKrSKlFUKV0WkcCXiAOcj4Ox2CLimbIeq/LKnQ1I0Z44cZtu6CA7uOkxa7GHqVotk/V89efWHZ2jSxASpfv1M5cGAgOJ7+6Qk+HPHOdrur4277TwvbvyV8JM9SU01pZEvfszrWEoKRERk3bNuXbj3XrNlDueJPwBLukHySagzFHovBJeLEtgfU+GP501oHrIDKtUv0udavhz+9S8ICzPPg4LgP/+BO+4wbZ8wwfw1HEzv1Wefgf9l6p4U2ol1sGqgmUPX4mlTvKOQC2eXK2mJsLgNxP9t/kCSEGmWaPANhuu2Zg6DdVoJR0ywiv8bKjeG/mvA5yozb3Pv2xD2L7CnmuGwPedBzZ6ObnGpmjoVnn/e/LcsPByq5ZEvU1Phv/+Fl14yve61a5thgoMHl257RfJL4aqIFK5EpCScPAk//WQ6dfr2NQsll7jf7jHrXdXsBb2+Ba+CJbh9++Cjj+CTT8zcJjALPV9/PTx2/xmuTe+B7dxe8O8A/deCe+XsN7CnwbJeZv5VwDVw7YpCzb/ZudOUcl6yxDz384N//xseewy8L8k0n35qetqSkqBxY1O1rG1xToU5uwOW94bUuKxjlRuZ9d7qXFeMb+SEdj4Pu6aCdx24PtwsNr64jRnu2+xR6PSOo1tYeAnHTA/tuX3mn3e/1VDpkjUxz2w1QyLj95v5jm1egJaTyuecs0vs2QPt2pk/ysybB7feigmaKWfz/O/Oli0werR5LcADD5h5k5Ur53q5iMMUJBuU5IAJERG5SM2aMHasWfulVIIVmN4DF084uQ4Wt4VjSwr08qZN4bXX4OhRM5epd2/zl+ZfFifh/tsIbOf2EpsWxMlWP+YMVmB6sXp8Dm6Vzbpne94o0PsfOQL33GOG/S1ZYuaBPfGEKV7x9NM5gxWYn++GDVkFQLp1gy+La8pX/AEz5C01zgTWHvPA+ypzfPVg+HWk+RJeEcX9BbtfMfshb5m16LxqQLc55thf78Kxnx3WvCJJjIIVfU2wqtQA+q3KGazALK0weBs0uNMMh905BVYNKPe/E3Y73H+/CVZDhpihflgWrLkBvrsqz3/unTrBtm0wfrx5/sEHJqCtX19qTRcpdgpXIiLlmX87uG4z+LUy616tvg62TjCLLBeApyfcfjusWQN/7rKz4Y276R28jtgEX3pOXkydxnUYOdIU58gxHqJKYwi5UJ5957NwZtsV3y8uzqyN06wZzJ5t7jlypBlqNGOGWcj5cjp0MH8VHzjQzEm74w7zBS41tUAfO7vEaFg5wPwcq7YzJe4b3Gp6aJpPMD0VEd/Aj8FmiJg97zl35Y5lweaHzfptta+DoJuzztUZBM0eM/u/3QNJpxzTxsJKPA4rrjULY/vUuxCsLvPXEfcq0OMz6PapKc9+fBX83A6OLi69NpeyTz4x5dZ9fMywXJsN+PtjiPoFrDT4baz5OebC2xumT4cVK8ww3wMHzB9xnnnGzOEScTYaFpgLDQsUkXInLRHCnja9B2DCQc8vTansggr7N+yehmVzY2nqL7wwqx+//ZZ1umlTU3a5Xj3zxcnLC7y9LNrG3UL1hAUkewVzvP1WvCr7mHPeplCGzWbCzwcfwIsvwqkL38GvvtoMFerateBNTU+H554z8zvAVDH7+muoVauAN0qJMXNtYnaQ5tWIr86sZ8HiWiQmQvPmEBwMnRqH0Tb5QTziLlTV8O8And+HGvlbBuRSCQnw118mUGZs+/eDry80apRzCwhwYJG6Q/Ngw+3g6gVDdplAfbG0RPglBOLCoe4IU549l8aeP28qQO7ebX4Hsi1j4AhJJ0yPVexuM4+q/2ozJDC/4vbC+lvhbJh5HvwktJsGrh4l0VqHiI6GFi3M2lZvvmnmPZJwDH5qCamxptc6Ld4sbn7Nj5f9JY2NNT3Tn35qnl97rVl8OLceapHSpDlXRaRwJSLl1tEf4be7IfmU+SLccTo0eSD/38r3fwi/P2D2u82BRncBpjz8Bx+YKn7nzuX+0uqVT7HzlbbU8Y/inSWP8vjcrPk3Li7mC5TNBvHx5ljz5vDqq3DDDUUPDQsXmgpl586Z0vLffgv5XaHDSk0k/sdBVElcx6nzgXR9dj0HTjTO9Vqbzc7jgz9i6s3P4OsVg2XZOOj2ICkt/kvD5lXx9Mz5mtOnTXDasyd7kDp8OJdewMvw8ck9dDVqZIZIltgX1JQY01uXdBzavgStn839ujPbYWlXMw+n68fQ+J7MU6mpprLciy+aNZEAqlQxVeTuvNNBoTHppOmxit1lhn72Xw1VCrGIWnqyWdQ7Y3Htap1MsYvC3KsMuvVWCA016+v99tuFiqJrb4Ij35nP2vX/YElXsCdDp3eh2SNXvOd335l/X+PjTe/z99+bP9KIOIrCVREpXIlIuZYYDRvvguil5nnd4dDl/8z8mMs59jOsGWbmkrR+Htq+kOOS+Hgzmf2nn8x+YqIpLJHx2LXeEubdb4o+DH5tMb/syFkeLCDAfMkeN87MsSoue/fCTTeZXhF3d3jrLVP4Ircv7gkJZpjSTz+mcWONmxnUahGxCb5c8/Iadka0p0sXGDbMtHXvXhOM9u41Q5rsdgjwPc7rt/+LMb0+AyA6JpCnvnyTTcdvIzjYRkCA6YUKDzeFTvJSrZrpFcjYmjY1P9cDB7JvkZFXDmItW8JTT5mwUpw/VzY/Avtmgm9zGLwDXHNJkBl2vwZhE81wucE7sFdqzDffmLL6+/ebSxo3NvMTM3pDb70VZs0y6yKVmuTTJljF7DQLcPdbA75Ni3bPI4vMHzZSzpjenC4fFNvyBI7y00+muI2rK2zebIbjEjEffr0FbG6mQqR/W9j7Dmx93PxBZ9AWqNrqivdet84sA5GQYCp/LlhgFlMXcQSFqyJSuBKRcs+yw963IOwZM0/GuzZ0nwu1+ud+/dkwU/UvLR4ajjG9VoXtTtjyBPz1NpZXIEnX/kGivWZmAEtOhoYNS66X5dw5UyDj22/N87vuMl/cvb1N8YyffoIffjDBKinJ4pP77+Hua+aQmOLFfzYsoXH33gwZAoGBud8/OdmEhD17zOZychW3NXmIBtX2ArB8Vz8enj2TfdHNsr0uKCh7iAoONo81a+bvx5ycbMrmXxq6Mra4iwob1q9v5rPcfTe59qQVyOnNplcCy1SCrHXt5a+3p8PKfnBiDTFu3Rj0+jp+32JK9wcEmCGc991nvqy/8oop652eboaYfvaZmYtzWekpZl2tonR1JZ8xbTwbBl6Bpty6b/MrvixfEo7A+ttNgRkwvcad3s2+fIGTiI83gT0y0iyP8NprmJ/dTy1NL2arZ6HdS+Ziy4LVQ8wcrKptYdDvlw/hF6xaZQpkJCXBjTeaHrJi/cOASD4pXBWRwpWIVBhnw8yXvbhw87zFU9D2P9nnhJyPhKXdIPEYBF4LfX4u2pyRtERY0hli/zS9Zr2+K9VxX5YFb7wBEyeaXqbWrc0Xtu3bs1/3/gP/4oHe/8OOK+ndF+De8IbCvWF6Mlb4/2DXy9jsSaRbHvx6ZiLR/k/RONiX4OCSLT1tWWb+2qefmrlrGcPu6tQxFRfvu88MKSwwezos6QJnL1TH6/FZvl6287cIGoW3pbJnLM99+yLTlz3H00+buTqX/hw2bTLFSP7+2wwdnTTJBK4cX7DTzsOOyfDXe2advOrdoGZ3qNEdqnXOvZJlblLOwor+5jN5BZhy634t8vfa/LKnwa6X4c+XzB85ag+Gq7/OfxvLiAkTTHGZhg3hjz+gUiWyln7wDYbBYdkDVGK0qViafNIUgAl5M1/vs3Sp6SVOSTFFbb74Io/FzEVKkMJVESlciUiFkpYA2/4J+983z/07mDkhvs0hJRaW94KYP0zFwQG/Fs8iz2d3mC/m9hTo8hE0ubfo9yyglStNyeiMwhk2mynbfv31cHe316kd/bQ50W02NBpb9DeMPwCbH4WoC2Wp3X1Nz0XzJ8xCtKUgMdEs1vrqq6anDkzv2D//CQ8/bOY55VvGUC/3qnD9HvDOozvvgv37zfC/0FC4rceXfPnIHaRbrsR1XY9/k7yrlZw7Z4oczJ5tnnfpYr5gN8mYsnR8DWy6x/x8c2NzMb0lNbpnbZUb5wz0KTGwciCc2QyeNUywysfwtUI78j2svw3SE8G/I/T5CbwLWmnFMTZvNv+u2O3wyy8waBAQvdxU08QGA9blvoDy0R/N0GKAvkug9sB8vd9PP5meq9RUE7Y//dT0boqUFoWrIlK4EpEKKXIh/H6vmW/i6gMd34TI+RC9DLxqwaDfoFL94nu/8P/B9n+Z9xocVvg5LZZlig4knYSaPcy8jnyKjIR33jHDm4YMMUPT+Hu2+bIO0OF105tXXCwLIheYkvRxF1ZOtbmZuTctnoKqbYrvvS4jORnmzoVp0+DgQXPM39+Uq3/sMbN/WQnHTBGLtHPQeRY0fTDPS6Oj4aWX4MMPIS3NZJo77oBZd95O5dPzoHITGLz9ij0333xj1lKKiTG9JLPeiefO1pOw7btQAdOnrqnO6FkDTm3M2hIic97Ms2Zm0LJqdOevqOYE7h1B1fRNJFOdn1NWEZXYJtt8wUv3M557eJiFcG+4oRBf+E9tgjXXmwIzlRqYXmG/4ALepHSlpkLnzqaIzR13mCI2pJ2Hn9rA+YNXXiw6Y46eVy0Y8seV53pe8P33cMst5nfo7rtNARQXZ11QKD0ZIr421VtT40wVxUsrbEqZonBVRApXIlJhJRyDjWPg+IqsY26VoP9aqNaxeN/LssPK/mYdoOpdTK+YSz4nVNjT4dQGOLLQbBm9Fm5V4KphUO8fUHsQuBVw8taR72HdTaZtLZ6GDq8W7PX5Zdnh2GIIfx1OrM06Xvs6E7ICry2VoZJpaaYAyX/+YwpygOm9evRRM+yrZs08XvjrrRARCtW7wsANpnfoEnFx8Prrpjx3QoI5NniwCXTt2mGG4C1uZ8JP4/ug64dXbG9kpAkytpOr+Pi+cTQKuJAMm9xvgrB7Lv/PTjhqQtbJDebx7DbTY5qL0+eqce1/V7Izot0V23Kppk3hySfNPL4CzRk8tx9WXQfxf4OHP/ReBAFX53ppSopZTDsuzoS5AvU0FpPXXzfDSatVMwVZAgIwPd973jTl6of+adb6yku2svwFGxb87bemwEl6ugna77/vwOUHCiMxGvbNMqMEkk5kHa/SDAZuBM9qJfv+KTGmmJHN1SwLkMfvmeSkcFVEClciUqFZdgh/A3ZONpUBe38PV11fMu91PtLMw0iNgdbPQdsX8742LdGEviMLTeW15IvK7Ll4mi+mSdFZx9wqm3YH3QJ1BoPbFSYWHV8DqwaZktGN7jElpEvjm9up32HP/0wvoWU3x/w7mpBV7x+lUuwgPR3mz4eXXzbzZ8DMw3rgAVOswM/PlIw/fRqsY0vpEDMIu+XCx0e3sDuqQ+a5i7eYmKwKhl26mKGIffpc8sbHV8GKfoAFvReaL9uXk3oO+7aJuPw9C4DDp+oxaeH/cf9zA3Le+yInT5phoCtWwLrVyVS1ttG96Ua6N91Ij6YbuKraMWIT/Xli0Qoi4zvg5UXmGmwZ+5c+z9jfv998yT971rxXzZomnD788JUXu86UdNIMlzu9yfwu9/gc6t0CmJ/h1q2mp3HevKxhrJUrm4W9H3gAOhbz3z3ycuCAmaOYmAhz5pggyanfYVl387vbZ7H5d+1KzoZdKM+eYqomNrk/32348ktT8dKyzM/57bcdG7D27DHFVhISzL8nuW0BbpsJOPsWXse/xmZdWMnc+yozLPjvj8wfGAJ6Q9+l+Sr0UShpCbBqIJxcn3WsRg9oOdH8dzKXP5BIFoWrIlK4EhHB9AalJZbsvBOAw6FmoVWbC/T/1RQiyJByFo7+ZAJV1C9m+FEG96rmS0HdERd6qXzg9O8Q8Q1EfAsJEVnXuvrAVUNN0LpqqOmNu9iZ7bCijxmiU3cEXP1N6VdwO/c37JkOBz4x83DADMNsPgEajyuVggd2u1m09aWXYMuWnOe93BP545U2NKn1N9N/Hs+Tn0+/7P2Cg02v2I03XuYL8PanTQ+eZw0zTCyveUfRK2DTODh/GICTfg8y6JnX2L6rCjab6U2ZOtUM0zt3DtauNWFqxQrYuTP7rVxdzaLU/fpBv2sturU7iqePT6F7DuLj4ZNPYPp0OHTIHPP2NpUpn3wyn4shpyWYhZiPfA/YiGn8Jh+sHs/cuWb5gAy1apmFpP/6K+tY584mZN1664XCEiXAsszcqmXLzOK+y5eDzZ5ieqFidxWoqAkA4W/C9n+Cqzdct61AwyE//dQMDbQs8/P93/9KN2ClpcGiRWYdtpUrc7/GzTWVmzvP5/FBb9Oj2cbM4xv29eCT9U+w/vCNVKrizjXt/+DVfj1xs86Zn2H3ucX/YeypsHaE6S13r2r+iHF4XlYPrl9L01Nf/7ZytcB1cVK4KiKFKxGRUrZhNBz6HCo3gmt+gOiVJlCdWG16zzL41DXhp+4I85fevIYRWpYpEx75rQlb5w9lnXP1Nn9dD/qHCVpJx2FZTzNMJ+Aa6PtLgeZtFbvk0/DXTPjrnazeOfeq0PQhaP6YKZtfwizLVGl76SVYf+EP3e7uMO325/nnwKmcPF+HJ5btoZJfFapXJ88tX6Xk05NNL0bMDlM5r89P2V+UGmcC2P4PzPNKDcwixLWu5fx5M3zxo4/MqXbtTI/Opk3mC/DF2ra9EKb6mZLuJTGkLi3NDF17/XXYts0cc3GBm282PYCdO1/+9efj0zn6/eM0s80EYPrP4/nnF2/g6enCjTeahXX79zfhcO1a02M2f76ZBwUmdN15pwlabdsW72f7/HMzJNPLy/RuNmkC/PES/PGcCcZDw/M9fwowPV2rBplCGP4dzbC4Anyx/+gjMzQQzNIC//1vyQes6Gjzvh98AEePmmMuLqYITosWEBsL9sST9Az4kMFNZlKz8jEAUtLc+Wrjrby95HG2HuyU474D2ixl8dNDcHNJJ7X5c7iHXKYHv6Asu/nv6+EvzX/7rl1mio0kRsGeGWaIYuqFtRp86prhgo3vc7rqlSVN4aqIFK5EREpZSiz83C6zVyIbv9YmTAWNMF/CCvoNyrLMPJuIC0Er/u+scy6e5ktE8mnwb28qxHn4FfpjFKu0RDj0mRmiee5CN4WLB9ToBr4tzF+bfVuYUuHeV5XIN0vLMkPqvL2hsv0vbD+3MX/tvvqbzGFrxSLmT9MDYk+GTu9Bs4fN8ailsOnerKIUTR+B9q/k+OL33Xdw771w5kzWsUaNssJU374X5gaVEssyazT973/w889Zx3v3NiFryJCsYgx2uwlKc+eaoh3x8Rb/uv51XrttIgAH0m6h+vWf4Vct98B/8qQZovfBB6ZkfYbu3U3IGjmy6OvGnTplwsOpUybETJoExO6GnzuY34ceX0KD2wp+44RjsLiNWVi55UTzz7YA3nvPDA0EU6L/hRcK3oQrsSz49VfzXvPnZ4X2mjXN79wDD5i14zgbBnvfhkNfmt9jMOukNX0ImjyA3bMW586ZAHbxtmSJqYQ5qtP/8X/33QdAaMQcut9+F/XqFUPjtzwG+94zhXOu+QHqXJf9mpRYE7D2zMgaVu3hb/5da/44eOU18bIUpKdA3G7z32YHU7gqIoUrEREHOLHOLN5qTzN/Wa07wgxfqdLkii/NN8syPSQR35jt3D5zvHITU1DjCuXEHcKymzlm4a+bIh65cati1hbya5EVuHxbmJ7A4hjeaFmmzPbxFaboRp/FxR/m9r4NW58wvYb9VsPf/2c2gEoNodsnENgnz5cfPWpCRq1aJlA1aFC8zSusXbtMyPryy6wephYtTGXGyEgzX+fwRX9TaNTI9FA9OHgegQfuMkO6al5t5j5eZtii3W6GqH3wASxcmBUCqlY1c6MeeMC8b24sy1SQPHcu5xYXZ9q+aBG0aWPmf7m7ppslGk5thDrXwzWLCv/7ELkQ1t0I2KDfCgjsW6CXT59uhgaCGYL6738XrhmXio83vXUzZ2bNQwQTWh95xFQu9PQEYnaZAHNiddZF1TqZJRbqjcxXb1xsrAlYnnsm8VDvV0hJc2fwa0vwD+7LE0/A1VcX8se783nYNRWwQY8vLh+A05Pg4Gew+zWI32+OuXpBo3HQ4p9QuWEhGlAI9jQzFzMi1FRWtafATSeuPGe2hClcFZHClYiIg5yPMP9D9yqFboaMEu7H10DQTeBTp+Tfs6hi/jB/IY8NN9XW4sJNtbmLh05ezMUDqjQ1Qatqa6gWYr74FXQ9pUPzzHwgVy8YsqtkykZbdlg1GKKXZj/e7HFo/9+c8+SczNGj8NZbJvzExWU/5+trepjuugt69rzoi/Tx1WauTGqsWXeuz8/5+pIbHW3mgH30UdYcMDDzzHx9cw9Rlw6jvJTNBhs3mntkrnHmVgWu322GkxXF7w/A/g9ND+yQnQWe+/baa2ZRcDBDMp8qwuoJ4eEwa5aZ15Xxz8nb25Sdf/hh6NDhwoX2VPjzFbMYtD3V9AzVu8X8vtboVqg0lJ5mJ3rB7VyVFsrZ81Xp8cIG9hxrQYcOZq23W2+9EOjyI+OPFcBOz/f4ef/DhIebqqDu7tCsmdmaNzePjRub+YrY0+HId7D7VThzYeKlzdUExRb/NOsgFnfxC3s6nFxr5t9GzjdLE2TwCjS/99U65P36UqBwVUQKVyIi4jTSU8xfmjMCV2bw2pNVGONS3ldlBa3qncx+XoE2JcasaZV0HNq+DK0nl9hHyTZMrHIT01sV0Kvk3s8B4uJM6Pn8c6hd2/RSDR9+maF7MX/C6sFmaKRXoJmTVi0kX++Vnm4KULz/Pvzwg+nduhIfHzMf7dJtxAgzDI7zh+GnVqa4TOeZZthbUaWdh587muGvQTebYacFDCcvvwxTppj9Rx4xP9v0dLPZ7fl73LcPVq/OumfTpiZQ3XXXJWu/ndkOv91tesHBLP/Q6V2oVNRxfJgepBX94NQGTiY2JGTSb0SeNP9uBgSYHsiHHjKfL0NSkilwEh5uturnPuexkNEATPlmKi8vnHLFt3VxgYYNLw5cFl0brKKF9SreMRf9wcPdz/z+Ve8M1Tpj9+9MXFoQsXE2YmJML9ylj3Fx4OZmfsd9fMzm7W2nnvd6Grp8TZ30b/G0Z1V6TXerQWqtm7E1GIXHVb2xlYEVoxWuikjhSkREnJ5lNz2BGYErZof5S3RsOJDL//p9grLCln/IhcBVAzY/auZs+DaHwTtKrlR0hpg/TLnohmMcPhSozEg4CquHQMxO04MX8o7pfbSnmC09JWvfnpzr8fi4ZA5H2Dhp601cpcH4+FbC1zd7gKpc+QoLIVuWaUfUL1CzF/RfXXy9GGe2wpJuYKVB10+g8d0FvsWUKSZkFYWLCwwbZkJV//6XLFScngy7XoLdr5jeYo9qZsHk+rcV7zDZpFOwtBvE/02qX1fe2rWKt9715sgRc9rd3axzlpxswtTBg1nBeWiHH1k4YQRurunM+PkJ/j1/Os2b22jRwgwLDQ42vZR//WV6sf76y2znzuXdnB7B25hyy2v0bfo9nm5JOc4fjw1g84HObD7QmS0HOrH5QGdOxuX2xxqLrk02MapbKP/o+g11qx3NPHMm3p8Fm28i9LdRrNrdl3R71nDmyEioW8TO0aJSuCoihSsRESm3UuPN0MIzWy5sWyFuL7kGrkr1TUDDgn4rCzwfRopRahysuwWilxX9XpkVMy8sTZDb4su5Ofg5bBxtCsEM2WECd3Ha/SqEPWMC5OCwAs+3tCwznG/dOhMSXVwK9lipEtx004UCFZc69Ttsugdi/zTPg24xvVUlNU8z7i8TsFLOQtDNpHb9mu8WuvDWW7Ahl6mX/v5w67VrmTF8EB6uSRxxH01qyBzqN3DJHhBzYVlmKGlG0Lo4eP39d9aQUTfXVFrV/ZPOjTZnbm2C/sDNNeew5OPx9TgY25mjSZ2JsVrSwGct7at9TXWvrCUy4lN8Wb3/Rn76YySrwvsTe86DxESzZljG/EQwhVSqVy/MD7H4KFwVkcKViIhUKKlxZqhTRtg6syWr2AcUfA0jKRn2VAibZOaluXjkvbl65n487Rwc/cGsYZfBxcOsExd0M9S9wVSKy03SCfixhRmy2e6/0GpSCXy+dFjZ3xSHqN7FFJnJa7mF0pKWCH88D3veML3BnjXNcMjirJaZlxNrTSEZe4pZVLzD64BZg+7HHyEwkMweqQC37dhW9jH/Ll81DHrNL5afXVqa6Rn76y8TwHx9zcLIVateeKySSFV24Hlus1n+4szmvP9YAxcWd78B6o8yv3d59ISnpZnFqhMTzWLcVwqIJU3hqogUrkREpMJLiYEz28xizPVGaoheeWFZpucycr5ZBy5ub9Y5mxvU6md6ZeoOz16Ge/1tcPgrUxZ70O8lF3rOR8LitpAaY96r9mDTppo9S3/9uZMbzNyqjKUQ6t8OIW8VbD2vojr4BWy80+x3ngVNH8x5Tdw+WH71hbX6ekOfX8CtiPX3iyI1zvyR5vQWE7Zi/zRLatQfZf55OrJthaRwVUQKVyIiIlLuWZZZryryWxO2Yi6qOW5zgYA+pkfLzceEDJurCVbVOpZsuyK/g/WjTE9dBhdPE7Bq9YPAfmZOYHEsM5CbtPOwY7KpuIdlFu7u/L7p2XOEjMWabS5wzY9mSGeGhCOw7GpTaMS/A/RbVXbW6itHFK6KSOFKREREKpy4v0zIivjWLLx9qRZPQ4dXS6ctCUchejlErzDrqyUey37e3deEv4yw5deyeIpKHF8Nm8ZlDZ1sdDd0fCPv4ZKlwbLMfK8Dc8ywugG/gn87s/j58t4mIFdpao6XxjIWFZDCVREpXImIiEiFFn/wQtCaD6d/A79WMGizY4Z0WZYZvnh8xYWwtcoMG7yYVy0IvBZqXWvaalmA3cyTwjKPlj3r2MXnM45HLYF9s8z9fIKgy4dQ57rS/KR5S08xJfmPrzRLKfRbARvHwOnfzfOB600BGikRCldFpHAlIiIickHSKTM0sKzMu7Onw9ntWWHr5K95r+lWGE0egA6v5b+KYmlJiYGlPczyCjY3U7beoxoMWGd67qTEKFwVkcKViIiIiJNIT4ZTGy/0aq00QwhtLoCLeczYcLkwdPDSYxc2d18z9LHWtY79PJcTfwiWdjXFK9wqwbUroUYXR7eq3FO4KiKFKxEREREpk86GmTXBmj4CAVc7ujUVQkGyQQmVWRERERERkWLn3x56znN0KyQPDl6SS0REREREpHxQuBIRERERESkGClciIiIiIiLFQOFKRERERESkGChciYiIiIiIFAOFKxERERERkWKgcCUiIiIiIlIMFK5ERERERESKgcKViIiIiIhIMVC4EhERERERKQYKVyIiIiIiIsVA4UpERERERKQYKFyJiIiIiIgUA4UrERERERGRYqBwJSIiIiIiUgwUrkRERERERIqBwpWIiIiIiEgxULgSEREREREpBm6ObkBZZFkWAHFxcQ5uiYiIiIiIOFJGJsjICJejcJWLc+fOARAUFOTgloiIiIiISFlw7tw5/Pz8LnuNzcpPBKtg7HY7x44do0qVKthsNkc3h7i4OIKCgoiMjMTX19fRzREnot8dKQr9/khR6PdHikK/P1JYJfG7Y1kW586do06dOri4XH5WlXqucuHi4kLdunUd3YwcfH199R8YKRT97khR6PdHikK/P1IU+v2Rwiru350r9VhlUEELERERERGRYqBwJSIiIiIiUgwUrpyAp6cnzz//PJ6eno5uijgZ/e5IUej3R4pCvz9SFPr9kcJy9O+OClqIiIiIiIgUA/VciYiIiIiIFAOFKxERERERkWKgcCUiIiIiIlIMFK5ERERERESKgcJVGTdz5kwaNmyIl5cXISEhrFu3ztFNkjJo7dq1DBs2jDp16mCz2Vi4cGG285b1/+3df2hV9R/H8dep6653t4tsLnc3zZo5lZkb6CyvWZazsVtIllHZkmv9IaNtbIhglLKJw0l/GIW1sFKMjMmo1f4wc5VdyJCmdvMylhhJCrZuEtW8sUnu8/0jOnTZT/ted67z+YAD93w+5969D7w4u++dHzOqr69Xbm6uPB6P7rvvPnV2djpTLFJKY2OjFi5cKJ/PpylTpmjlypU6depUwjbkB0NpampSYWGh/c86A4GAPv74Y3ue7GC0GhsbZVmWamtr7THyg+HU19fLsqyExe/32/NO5YfmKoXt379ftbW1evHFF/XNN9/onnvuUTAY1NmzZ50uDSkmHo+rqKhIO3fuHHT+pZde0o4dO7Rz5051dHTI7/frgQceUE9PzxhXilQTDodVWVmpo0ePqr29XX/99ZdKS0sVj8ftbcgPhjJt2jRt375dx44d07Fjx7Rs2TI9/PDD9hcYsoPR6Ojo0K5du1RYWJgwTn4wkrlz5+qnn36yl2g0as85lh+DlHXnnXeaioqKhLE5c+aY559/3qGKcC2QZFpbW+31/v5+4/f7zfbt2+2x3t5eM2nSJPPGG284UCFSWSwWM5JMOBw2xpAfXLmMjAzz1ltvkR2MSk9Pj8nPzzft7e1m6dKlpqamxhjDsQcjq6urM0VFRYPOOZkfzlylqEuXLun48eMqLS1NGC8tLdVXX33lUFW4Fp05c0bd3d0JWXK73Vq6dClZwgC///67JCkzM1MS+cHoXb58Wc3NzYrH4woEAmQHo1JZWamHHnpIy5cvTxgnPxiN06dPKzc3V3l5eXryySf1ww8/SHI2P66r+un4zy5cuKDLly8rOzs7YTw7O1vd3d0OVYVr0T95GSxLP/74oxMlIUUZY7R+/XotWbJEd9xxhyTyg5FFo1EFAgH19vbqpptuUmtrqwoKCuwvMGQHQ2lubtaJEyfU0dExYI5jD0Zy11136Z133tGsWbP0888/q6GhQYsXL1ZnZ6ej+aG5SnGWZSWsG2MGjAGjQZYwkqqqKp08eVJffvnlgDnyg6HMnj1bkUhEv/32m95//32FQiGFw2F7nuxgMOfOnVNNTY0OHTqkiRMnDrkd+cFQgsGg/XrevHkKBAK6/fbbtXfvXi1atEiSM/nhssAUlZWVpRtvvHHAWapYLDagCweG88+Tc8gShlNdXa22tjYdPnxY06ZNs8fJD0aSlpammTNnqri4WI2NjSoqKtIrr7xCdjCs48ePKxaLacGCBXK5XHK5XAqHw3r11VflcrnsjJAfjJbX69W8efN0+vRpR48/NFcpKi0tTQsWLFB7e3vCeHt7uxYvXuxQVbgW5eXlye/3J2Tp0qVLCofDZAkyxqiqqkoffPCBPv/8c+Xl5SXMkx9cKWOM+vr6yA6GVVJSomg0qkgkYi/FxcUqLy9XJBLRjBkzyA+uSF9fn7q6upSTk+Po8YfLAlPY+vXrtWbNGhUXFysQCGjXrl06e/asKioqnC4NKebixYv6/vvv7fUzZ84oEokoMzNT06dPV21trbZt26b8/Hzl5+dr27ZtSk9P11NPPeVg1UgFlZWVeu+99/TRRx/J5/PZf+WbNGmSPB6P/X9nyA8G88ILLygYDOqWW25RT0+Pmpub9cUXX+jgwYNkB8Py+Xz2vZ3/8Hq9mjx5sj1OfjCcDRs2aMWKFZo+fbpisZgaGhr0xx9/KBQKOXv8uarPIsT/7bXXXjO33nqrSUtLM/Pnz7cfjwz82+HDh42kAUsoFDLG/P1I0rq6OuP3+43b7Tb33nuviUajzhaNlDBYbiSZPXv22NuQHwzl2WeftX9H3XzzzaakpMQcOnTInic7uBL/fhS7MeQHw3viiSdMTk6OmTBhgsnNzTWPPvqo6ezstOedyo9ljDFXt30DAAAAgPGPe64AAAAAIAlorgAAAAAgCWiuAAAAACAJaK4AAAAAIAlorgAAAAAgCWiuAAAAACAJaK4AAAAAIAlorgAAAAAgCWiuAABIMsuy9OGHHzpdBgBgjNFcAQDGlbVr18qyrAFLWVmZ06UBAMY5l9MFAACQbGVlZdqzZ0/CmNvtdqgaAMD1gjNXAIBxx+12y+/3JywZGRmS/r5kr6mpScFgUB6PR3l5eWppaUl4fzQa1bJly+TxeDR58mStW7dOFy9eTNhm9+7dmjt3rtxut3JyclRVVZUwf+HCBT3yyCNKT09Xfn6+2traru5OAwAcR3MFALjubN68WatWrdK3336rp59+WqtXr1ZXV5ck6c8//1RZWZkyMjLU0dGhlpYWffrppwnNU1NTkyorK7Vu3TpFo1G1tbVp5syZCT9jy5Ytevzxx3Xy5Ek9+OCDKi8v16+//jqm+wkAGFuWMcY4XQQAAMmydu1avfvuu5o4cWLC+MaNG7V582ZZlqWKigo1NTXZc4sWLdL8+fP1+uuv680339TGjRt17tw5eb1eSdKBAwe0YsUKnT9/XtnZ2Zo6daqeeeYZNTQ0DFqDZVnatGmTtm7dKkmKx+Py+Xw6cOAA934BwDjGPVcAgHHn/vvvT2ieJCkzM9N+HQgEEuYCgYAikYgkqaurS0VFRXZjJUl33323+vv7derUKVmWpfPnz6ukpGTYGgoLC+3XXq9XPp9PsVjsv+4SAOAaQHMFABh3vF7vgMv0RmJZliTJGGO/Hmwbj8czqs+bMGHCgPf29/dfUU0AgGsL91wBAK47R48eHbA+Z84cSVJBQYEikYji8bg9f+TIEd1www2aNWuWfD6fbrvtNn322WdjWjMAIPVx5goAMO709fWpu7s7YczlcikrK0uS1NLSouLiYi1ZskT79u3T119/rbfffluSVF5errq6OoVCIdXX1+uXX35RdXW11qxZo+zsbElSfX29KioqNGXKFAWDQfX09OjIkSOqrq4e2x0FAKQUmisAwLhz8OBB5eTkJIzNnj1b3333naS/n+TX3Nys5557Tn6/X/v27VNBQYEkKT09XZ988olqamq0cOFCpaena9WqVdqxY4f9WaFQSL29vXr55Ze1YcMGZWVl6bHHHhu7HQQApCSeFggAuK5YlqXW1latXLnS6VIAAOMM91wBAAAAQBLQXAEAAABAEnDPFQDgusLV8ACAq4UzVwAAAACQBDRXAAAAAJAENFcAAAAAkAQ0VwAAAACQBDRXAAAAAJAENFcAAAAAkAQ0VwAAAACQBDRXAAAAAJAE/wOfeHKrXxqxqQAAAABJRU5ErkJggg==",
      "text/plain": [
       "<Figure size 1000x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model = Sequential()\n",
    "model.add(LSTM((50),batch_input_shape=(None, look_back, 1),return_sequences=False))\n",
    "model.compile(loss='mean_absolute_error',optimizer='adam',metrics=['accuracy'])\n",
    "history = model.fit(X_train, y_train, epochs= 50, batch_size=500, verbose=2, validation_data=(X_val, y_val))\n",
    "\n",
    "# Access the loss values from the training history\n",
    "train_loss = history.history['loss']\n",
    "val_loss = history.history['val_loss']\n",
    "\n",
    "# Plot the loss vs. epoch curve\n",
    "plt.figure(figsize=(10, 6))\n",
    "plt.plot(train_loss, label='Training Loss', color='blue')\n",
    "plt.plot(val_loss, label='Validation Loss', color='orange')\n",
    "plt.title('Loss vs. Epoch')\n",
    "plt.xlabel('Epoch')\n",
    "plt.ylabel('Loss')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "99cb9791-ce19-4ac1-b235-b0732c843f93",
   "metadata": {},
   "source": [
    "### Step 4: Model Evaluation"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "49b0b702-4dd1-4529-91b5-23a342750f98",
   "metadata": {},
   "source": [
    "#### 4a. Visualize predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "6c8a1601",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2/2 [==============================] - 1s 8ms/step\n"
     ]
    }
   ],
   "source": [
    "no_of_test_samples = 50\n",
    "predict = model.predict(X_test[0:no_of_test_samples])    \n",
    "final_predictions = [np.mean(x) for x in predict]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "01a648ce",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x1ecef6e7cd0>"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.scatter(range(no_of_test_samples),y_test[0:no_of_test_samples],c='red')\n",
    "plt.scatter(range(no_of_test_samples), final_predictions,c='green')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a35d8fc4-3f84-4d94-b805-1c3147cc7efe",
   "metadata": {},
   "source": [
    "### 4b. Calculate performance metrics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "1d955c49",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "219/219 [==============================] - 1s 7ms/step\n",
      "Mean Absolute Error (MAE): 0.03040154278278351 MWh\n",
      "Mean Squared Error (MSE): 0.0013240783009678125 MWh\n",
      "Root Mean Squared Error (RMSE): 0.036387886852025986 MWh\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import mean_squared_error, mean_absolute_error\n",
    "\n",
    "# Make predictions on the test set\n",
    "predictions = model.predict(X_test)\n",
    "predictions = [np.mean(x) for x in predictions]\n",
    "\n",
    "# Calculate Mean Squared Error (MSE)\n",
    "mse = mean_squared_error(y_test, predictions)\n",
    "\n",
    "# Optionally, calculate Root Mean Squared Error (RMSE)\n",
    "rmse = np.sqrt(mse)\n",
    "\n",
    "# Calculate Mean Absolute Error (MAE)\n",
    "mae = mean_absolute_error(y_test, predictions)\n",
    "\n",
    "print(f'Mean Absolute Error (MAE): {mae} MWh')\n",
    "print(f'Mean Squared Error (MSE): {mse} MWh')\n",
    "print(f'Root Mean Squared Error (RMSE): {rmse} MWh')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.18"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}