LSTM NCSU SKUs.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import keras\n",
    "import wget as wg\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# Setting seed for reproducability\n",
    "np.random.seed(1234)  \n",
    "PYTHONHASHSEED = 0\n",
    "from sklearn import preprocessing\n",
    "from sklearn.metrics import confusion_matrix, recall_score, precision_score\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Dropout, LSTM, Activation\n",
    "%matplotlib inline\n",
    "from decimal import Decimal\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")\n",
    "import pickle\n",
    "from keras.models import model_from_json"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['id',\n",
       " 'DayOfWeek',\n",
       " 'Date',\n",
       " 'Sales',\n",
       " 'Customers',\n",
       " 'Open',\n",
       " 'Promo',\n",
       " 'SchoolHoliday',\n",
       " 'High_Var',\n",
       " 'Luxury']"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X=pd.read_csv('~/Desktop/ML/ten_skus.csv')\n",
    "X=X.iloc[0:,:]\n",
    "list(X.columns.values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "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>Store</th>\n",
       "      <th>DayOfWeek</th>\n",
       "      <th>Date</th>\n",
       "      <th>Sales</th>\n",
       "      <th>Customers</th>\n",
       "      <th>Open</th>\n",
       "      <th>Promo</th>\n",
       "      <th>SchoolHoliday</th>\n",
       "      <th>High_Var</th>\n",
       "      <th>Luxury</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1/1/13</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>1/2/13</td>\n",
       "      <td>5530.0</td>\n",
       "      <td>668.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>1/3/13</td>\n",
       "      <td>4327.0</td>\n",
       "      <td>578.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>1/4/13</td>\n",
       "      <td>4486.0</td>\n",
       "      <td>619.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>1/5/13</td>\n",
       "      <td>4997.0</td>\n",
       "      <td>635.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Store  DayOfWeek    Date   Sales  Customers  Open  Promo  SchoolHoliday  \\\n",
       "0    1.0        2.0  1/1/13     0.0        0.0   0.0    0.0            1.0   \n",
       "1    1.0        3.0  1/2/13  5530.0      668.0   1.0    0.0            1.0   \n",
       "2    1.0        4.0  1/3/13  4327.0      578.0   1.0    0.0            1.0   \n",
       "3    1.0        5.0  1/4/13  4486.0      619.0   1.0    0.0            1.0   \n",
       "4    1.0        6.0  1/5/13  4997.0      635.0   1.0    0.0            1.0   \n",
       "\n",
       "   High_Var  Luxury  \n",
       "0       0.0     1.0  \n",
       "1       0.0     1.0  \n",
       "2       0.0     1.0  \n",
       "3       0.0     1.0  \n",
       "4       0.0     1.0  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "sequence_length=7\n",
    "\n",
    "def gen_sequence(id_df, seq_length, seq_cols):\n",
    "    data_array = id_df[seq_cols].values\n",
    "    num_elements = data_array.shape[0] #number of rows\n",
    "    for start, stop in zip(range(0, num_elements-seq_length), range(seq_length, num_elements)):\n",
    "        yield data_array[start:stop,:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[  1.00000000e+00,   5.00000000e+00,   3.70600000e+03,\n",
       "          4.59000000e+02,   1.00000000e+00,   0.00000000e+00,\n",
       "          0.00000000e+00,   0.00000000e+00,   1.00000000e+00],\n",
       "       [  1.00000000e+00,   6.00000000e+00,   4.36400000e+03,\n",
       "          5.00000000e+02,   1.00000000e+00,   0.00000000e+00,\n",
       "          0.00000000e+00,   0.00000000e+00,   1.00000000e+00],\n",
       "       [  1.00000000e+00,   7.00000000e+00,   0.00000000e+00,\n",
       "          0.00000000e+00,   0.00000000e+00,   0.00000000e+00,\n",
       "          0.00000000e+00,   0.00000000e+00,   1.00000000e+00],\n",
       "       [  1.00000000e+00,   1.00000000e+00,   6.10200000e+03,\n",
       "          6.12000000e+02,   1.00000000e+00,   1.00000000e+00,\n",
       "          1.00000000e+00,   0.00000000e+00,   1.00000000e+00],\n",
       "       [  1.00000000e+00,   2.00000000e+00,   5.01100000e+03,\n",
       "          5.60000000e+02,   1.00000000e+00,   1.00000000e+00,\n",
       "          1.00000000e+00,   0.00000000e+00,   1.00000000e+00],\n",
       "       [  1.00000000e+00,   3.00000000e+00,   4.78200000e+03,\n",
       "          5.23000000e+02,   1.00000000e+00,   1.00000000e+00,\n",
       "          1.00000000e+00,   0.00000000e+00,   1.00000000e+00],\n",
       "       [  1.00000000e+00,   4.00000000e+00,   5.02000000e+03,\n",
       "          5.46000000e+02,   1.00000000e+00,   1.00000000e+00,\n",
       "          1.00000000e+00,   0.00000000e+00,   1.00000000e+00]], dtype=float32)"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sequence_cols = ['id','DayOfWeek','Sales','Customers','Open','Promo','SchoolHoliday','High_Var','Luxury']\n",
    "# generator for the sequences\n",
    "seq_gen = (list(gen_sequence(X[X['id']==id], sequence_length, sequence_cols)) \n",
    "           for id in X['id'].unique())\n",
    "seq_array = np.concatenate(list(seq_gen)).astype(np.float32)\n",
    "seq_array[934]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "def gen_labels(id_df, seq_length, label):\n",
    "    data_array = id_df[label].values\n",
    "    num_elements = data_array.shape[0]\n",
    "    return data_array[seq_length:num_elements, :]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 5263.], dtype=float32)"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# generate labels\n",
    "label_gen = [gen_labels(X[X['id']==id], sequence_length, ['Sales']) \n",
    "             for id in X['id'].unique()]\n",
    "label_array = np.concatenate(label_gen).astype(np.float32)\n",
    "label_array[934]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 195,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[  2.00000000e+00,   0.00000000e+00,   0.00000000e+00,\n",
       "          0.00000000e+00,   1.00000000e+01,   1.00000000e+00],\n",
       "       [  3.00000000e+00,   6.04360000e+04,   1.00000000e+01,\n",
       "          0.00000000e+00,   1.00000000e+01,   2.00000000e+00],\n",
       "       [  4.00000000e+00,   5.28480000e+04,   1.00000000e+01,\n",
       "          0.00000000e+00,   8.00000000e+00,   3.00000000e+00]], dtype=float32)"
      ]
     },
     "execution_count": 195,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "seq_array[0]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "9350"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(seq_array)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "9350/9350 [==============================] - 19s 2ms/step\n",
      "Mean Abs Error: 2422.1649655853107\n",
      "MSE: 0.9841171621875722\n",
      "9350/9350 [==============================] - 1s 66us/step\n"
     ]
    }
   ],
   "source": [
    "nb_features = seq_array.shape[2]\n",
    "nb_out = label_array.shape[1]\n",
    "\n",
    "model = Sequential()\n",
    "\n",
    "#model.add(LSTM(input_shape=(sequence_length, nb_features),units=21,return_sequences=True))\n",
    "#model.add(Dropout(0.2))\n",
    "\n",
    "model.add(LSTM(input_shape=(sequence_length, nb_features),units=21,return_sequences=False))\n",
    "#model.add(Dense(units=2))\n",
    "\n",
    "model.add(Dense(units=1,activation='linear'))\n",
    "model.compile(loss=\"mse\", optimizer=\"sgd\", metrics=['mae','mse'])\n",
    "\n",
    "model.fit(seq_array, label_array, epochs=200, batch_size=20, verbose=0)\n",
    "# training metrics\n",
    "scores = model.evaluate(seq_array, label_array, verbose=1, batch_size=1)\n",
    "\n",
    "print('Mean Abs Error: {}'.format(scores[1],0)  )\n",
    "print('MSE: {}'.format(scores[2]/10000000,0)  )\n",
    "\n",
    "\n",
    "prediction=model.predict(seq_array, verbose=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x11a8e57f0>]"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11a697cf8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "y1=X.iloc[sequence_length:,].reset_index(drop=True)\n",
    "y1=y1.loc[:,'Sales']\n",
    "#y1.shape\n",
    "\n",
    "pred2 = pd.DataFrame(prediction)\n",
    "pred3=pd.concat([pred2,y1],axis=1)\n",
    "#pred3.head()\n",
    "\n",
    "plt.plot(pred3.iloc[:100,0]   ) \n",
    "plt.plot(pred3.iloc[:100,1]   ) \n",
    "\n",
    "#plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "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>0</th>\n",
       "      <th>Sales</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>4931.867188</td>\n",
       "      <td>5580</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>5322.408691</td>\n",
       "      <td>5471</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>5322.408691</td>\n",
       "      <td>4892</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>5322.408691</td>\n",
       "      <td>4881</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5322.408691</td>\n",
       "      <td>4952</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>5322.408691</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>7385.768555</td>\n",
       "      <td>4717</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>5322.408691</td>\n",
       "      <td>3900</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>5322.408691</td>\n",
       "      <td>4008</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>5322.408691</td>\n",
       "      <td>4044</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             0  Sales\n",
       "0  4931.867188   5580\n",
       "1  5322.408691   5471\n",
       "2  5322.408691   4892\n",
       "3  5322.408691   4881\n",
       "4  5322.408691   4952\n",
       "5  5322.408691      0\n",
       "6  7385.768555   4717\n",
       "7  5322.408691   3900\n",
       "8  5322.408691   4008\n",
       "9  5322.408691   4044"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred3.iloc[:10,:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 187,
   "metadata": {},
   "outputs": [],
   "source": [
    "#pred2.to_csv('~/Desktop/ML/predictions.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>DayOfWeek</th>\n",
       "      <th>Open</th>\n",
       "      <th>Promo</th>\n",
       "      <th>SchoolHoliday</th>\n",
       "      <th>lagSales</th>\n",
       "      <th>lagCust</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>5530.0</td>\n",
       "      <td>668.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>4327.0</td>\n",
       "      <td>578.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>6</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>4486.0</td>\n",
       "      <td>619.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>4997.0</td>\n",
       "      <td>635.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   id  DayOfWeek  Open  Promo  SchoolHoliday  lagSales  lagCust\n",
       "0   1          4     1      0              1    5530.0    668.0\n",
       "1   1          5     1      0              1    4327.0    578.0\n",
       "2   1          6     1      0              1    4486.0    619.0\n",
       "3   1          7     0      0              1    4997.0    635.0\n",
       "4   1          1     1      1              1       0.0      0.0"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# BUILD REGULAR NON-LSTM, RNN MODEL\n",
    "\n",
    "X=pd.read_csv('~/Desktop/ML/ten_skus.csv')\n",
    "X=X.iloc[0:40,:]\n",
    "\n",
    "\n",
    "y=X['Sales']\n",
    "Xs=X[['id','DayOfWeek','Open','Promo','SchoolHoliday']]\n",
    "Xs['lagSales'] = X['Sales'].shift(1)\n",
    "Xs['lagCust'] = X['Customers'].shift(1)\n",
    "#Xs['Mov_avg']=pd.rolling_mean(X['Sales'], window=7).shift(1)\n",
    "\n",
    "y=y[2:]\n",
    "\n",
    "y.reset_index(drop=True, inplace=True)\n",
    "Xs=Xs.iloc[2:,:]\n",
    "Xs.reset_index(drop=True, inplace=True)\n",
    "Xs.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    4327\n",
       "1    4486\n",
       "2    4997\n",
       "3       0\n",
       "4    7176\n",
       "Name: Sales, dtype: int64"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "38/38 [==============================] - 0s 3ms/step\n",
      "mean_absolute_error: 809\n",
      "mean_squared_error: 0.1\n"
     ]
    }
   ],
   "source": [
    "model = Sequential()\n",
    "model.add(Dense(20, input_dim=Xs.shape[1], activation='linear'))\n",
    "model.add(Dense(50, activation='linear'))\n",
    "model.add(Dense(30, activation='linear'))\n",
    "model.add(Dense(1, activation='linear'))\n",
    "# Compile model\n",
    "model.compile(loss='mse', optimizer='adam', metrics=['mae','mse'])\n",
    "# Fit the model\n",
    "model.fit(Xs, y, epochs=1000, batch_size=2, verbose=0)\n",
    "# evaluate the model\n",
    "scores = model.evaluate(Xs, y)\n",
    "#print(\"\\n%s: %.2f%%\" % (model.metrics_names[1], scores[1]*100))\n",
    "\n",
    "print(\"%s: %.0f\"%  (model.metrics_names[1], scores[1]))\n",
    "print(\"%s: %.1f\"%  (model.metrics_names[2], scores[2]/10000000))\n",
    "\n",
    "prediction2=model.predict(Xs, verbose=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MAE: 809\n",
      "MSE: 0.1\n"
     ]
    }
   ],
   "source": [
    "print('MAE: 809')\n",
    "print('MSE: 0.1')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saved model to disk\n"
     ]
    }
   ],
   "source": [
    "# serialize model to JSON\n",
    "model_json = model.to_json()\n",
    "with open(\"NCSU_DC.json\", \"w\") as json_file:\n",
    "    json_file.write(model_json)\n",
    "# serialize weights to HDF5\n",
    "model.save_weights(\"NCSU_DC.h5\")\n",
    "print(\"Saved model to disk\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loaded model from disk\n"
     ]
    }
   ],
   "source": [
    "# load json and create model\n",
    "json_file = open('NCSU_DC.json', 'r')\n",
    "loaded_model_json = json_file.read()\n",
    "json_file.close()\n",
    "loaded_model = model_from_json(loaded_model_json)\n",
    "# load weights into new model\n",
    "loaded_model.load_weights(\"NCSU_DC.h5\")\n",
    "print(\"Loaded model from disk\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "mean_absolute_error: 809\n",
      "mean_squared_error: 0.1\n"
     ]
    }
   ],
   "source": [
    "# evaluate loaded model on test data\n",
    "loaded_model.compile(loss='mse', optimizer='adam', metrics=['mae','mse'])\n",
    "score = loaded_model.evaluate(Xs, y, verbose=0)\n",
    "print(\"%s: %.0f\"%  (loaded_model.metrics_names[1], score[1]))\n",
    "print(\"%s: %.1f\"%  (loaded_model.metrics_names[2], score[2]/10000000))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x120203c88>]"
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYAAAAD8CAYAAAB+UHOxAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4xLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvAOZPmwAAIABJREFUeJzsvXecXHW9///8TJ/ZKds3nXRIQkKANHrvIEV6MSCCXrHgxYZ65YLt6tefKBYQDFUEIqAERCQEEBHSEBIgISQkQBLStmRndnr5/P74nDM726ZsZnb3bM7z8chjds+eM3NOZua8Pu8upJSYmJiYmOx/WAb7BExMTExMBgdTAExMTEz2U0wBMDExMdlPMQXAxMTEZD/FFAATExOT/RRTAExMTEz2U0wBMDExMdlPMQXAxMTEZD/FFAATExOT/RTbYJ9APurr6+X48eMH+zRMTExMDMUbb7zRLKVsKLTfkBaA8ePHs3r16sE+DRMTExNDIYT4qJj9TBeQiYmJyX6KKQAmJiYm+ymmAJiYmJjsp5gCYGJiYrKfYgqAiYmJyX6KKQAmJiYm+ymmAJiYmJjsp5gCUCmSMfjPQ5DJDPaZmJiYmPSKKQCV4oMXYcmX4KNXB/tMTExMTHqloAAIIQ4UQryV8y8ohLhRCFErhFgqhNioPdZo+wshxB1CiE1CiLVCiMNynmuhtv9GIcTCSl7YoBMPqcedbw/ueZiYmJj0QUEBkFJukFLOllLOBg4HIsBfgG8Dy6SUU4Bl2u8AZwBTtH/XA3cCCCFqgVuA+cA84BZdNIYlyYh6NAXAxMRkiFKqC+gk4AMp5UfAucAD2vYHgPO0n88FHpSK5UC1EGIkcBqwVErZKqVsA5YCp+/zFQxVTAEwMRmavPMEtG4Z7LMYEpQqAJcCj2g/N0kpd2g/7wSatJ9HA1tzjtmmbetr+/AkoQnAnvcgFR/cczExMVFICU9eDy/+cLDPZEhQtAAIIRzAp4A/d/+blFICshwnJIS4XgixWgixes+ePeV4ysFBtwAyKSUCJiYmg08qrr6TG5dCOjnYZzPolGIBnAH8R0q5S/t9l+baQXvcrW3fDozNOW6Mtq2v7V2QUt4tpZwjpZzT0FCwnfXQRRcAMN1AJiZDBf17GW+Hj14b3HMZApQiAJfR6f4BWALomTwLgadytn9GywZaALRrrqJ/AKcKIWq04O+p2rbhSSIM3iawV5kCYGIyVEiEO39+/7nBO48hQlECIISoAk4BnszZ/H/AKUKIjcDJ2u8AzwKbgU3APcAXAaSUrcAPgFXav9u0bcOTZBQcXmiaYQqAiclQQbcALDZ4728qJrAfU9REMCllGKjrtq0FlRXUfV8J3NDH89wL3Fv6aRqQZATsHhgxE97+s/qgCTHYZ2Visn+jWwCTToSNz6v4XOO0wT2nQcSsBK4UyQg4NAGIB2FvURPaTExMKoluARz8afW44dnBO5chgCkAlSKhWwCz1O873xnc8zExMelMz66dBKMOhQ1/H9zzGWRMAagUuguocRoIixkHMDEZCugWgMMDU8+AbauhY3f+Y4YxpgBUCt0F5PBA3WRTAExMhgK6ANg9cOAZgIT3h28yYiFMAagUugsIVBzAFAATk8FHDwI7qtT30j9mv3YDmQJQKZLdBKD9Y4i2De45mZjs7+RaAEIoK+CDF1Xa9n6IKQCVQncBgRIAyAaCMxnJO9vb+dOKj9kVjA3SCZqY7IckcgQA4MDTIRWFLa8M3jkNIkXVAZiUSCqh+o1kLQCVCfT2G69y3yofr7zfTHOHahB3wCseHr1+ASMD7sE6WxOT/YdkGGwusGhr3/HHqILNDc+yreEY/rx6G0++uY1qt4MfnX8ws8ZUD+75VhjTAqgESeVn/CQiuH3p+5z3wEZ2y2o2rHmNl97bzZGT6vjFxYdw/zVzaelIcPk9K0xLwMRkIMiNzQExaeOT+qNoe3MJx/5sGXe8uJEDaqvYHYpx/u9e46fPvUcsmR7EE64spgVQCTR/4h3/+oTHMhuZPbaaSO00zhLNnP+lU7BaOiuCH/jsXD6zaCWX3b2cR69fQKPfNVhnbWIy/ElGwFHFu5+0s3jVVv761iecGJ/I7Y5/8KN5KY45/jTG1Hhojyb58d/Wc+fLH/D8uzv52YWHcPgBw29+lSkAlUDzM0alg2e+fDQzRgXghQXw2m8gkwSLI7vr4QfUcv9n57Hw3pVcds9yHr3+CBp8zsE6c5NeiCRSXH3vKjJSMiLgYoTfpR61n5u0fw6baVAPeZIRtocFZ93xKg6bhdNnjOCSmdcin/g9lwXWQc35AATcdn564SzOmjWSm598mwvveo1rj5rATaceiNthHeSLKB+mAFQCzQUUxdnp2x8xU938mzd0BoU15o6v5b6r53L1fau4/J7lPHL9Auq9pggMFT5qibDyw1YmN3ppCSd4Yf0uYslMl32EgOOnNvDds6YzudE7SGdqUohMIkxzwsanDhnFbefOoNqjLcZWLlDpoCd+t8v+x05t4B9fO5b/+/t6/vDqFl5Yv4uffnoW8yfW9fLsxsMUgEqguYAiuPC5tP/ibEuIt3sIAMD8iXXce/Vcrrl/JVf+YQV/um4BtVWOHvuZDDyhWAqA/z1nBkdPqUdKSTCaYmcwpv61R9nSHOHh5R9x+i9fYeGR4/nKSVMIuO2DfOYm3UnHwkRxcvgBNZ03f1DpoEv/B/Z+DNXjuhzjddr44XkzOXPmSL79xNtccvdyLp4zhk8dMpp5E2oNbfkZ98yHMlqxibS5sFu1/+LaiSr4lKcg7IhJdSxaOJctzWGu+MMK2sKJgThbkwIEo2pylN+txFwIQcBj58ARPo6b2sAlc8fx7TMO4qVvHM9Fc8Zw77+3cMLPX+ZPKz4mndm/2w0PNTLxMBHp7CnOB56pHjf0PSPgyEn1PHfjMXz2qAn89a1PuHLRCg697Xm+8NAbLF61ld0h4yVymBZAJdCKTSyOHFeAxVrUbICjJtdzz2fm8LkHV3PlohX89NPKckhlJKl0hmRakspkSGo/T270MqnBdDlUklBcCYDPlX9FX+918pMLZnHF/AO47el1fOcvb/PH5R9xyznT+3QZRBIpPmyOsKU5jMtu4cSDGhFm2/CKIRNhotRlxTxL/WTVsmXDszD/+j6P9zhsfP+c6Xz9tKm8tqmFFzfs5sX1u3nu3Z0AzBoT4IQDGzl1RpOK/Q1xTAGoBJoLyOqq6rq96WB498mCswGOndrA3VcdzvUPvsHZv34170s5bBYeuGYeR0waHj7JoUgwqlxAfldxX5eDRwd47PMLeGbtDn7y7HouuXs5Z80aybmHjGJrW5TNezrY0hxmS3OYHe1dV423nDOda46aUPZrMNFIRogymhG9uecOPAOW3wWxILj8eZ/G47Bx8vQmTp7ehDxPsn5HiJc27GbZ+l3c8eJGfrVsIz84dwZXHTG+MtdRJkwBqASaC8jRXQBGzIQ37oP2rT38jN05/sBGnvnK0by/K4TNYsFhE9gsFmxWgd1qwaalkn7j8bVc9+BqHrluATPHDP0VhxEJxYqzAHIRQnDOIaM4eVoTv3/lA+765wf8be0OQAnJxAYvR0ysY0J9FRMaqhhfV8Udyzbyg2fWcUCdhxMPaqrItezvWJKR3l1AoNxAr/0aPlgGM84v+jmFEEwf5Wf6KD83nDCZ1nCCbz6+lu8veRe/2865s0eX8QrKiykAlUBzAdndvq7bcwPBBQQAYGqTj6lNvrz7PHTtPC6883UW3reSxZ8/wsxAqQDBWAqnzdKvYJ/bYeXGk6dy+bxxbG2LMKHeS43H3qub55eXzuaiu17ny396kye+eCQHjci/CjUpHWs6SgQn/t7EfMw8cNeqbKASBKA7tVUOfnP5oSy8dyU3LV6D32XnhIMa9+GsK4cZBK4EmgA4PN1uxk3TAVHWzqAjA27++Ln5WAR8ZtEKtu/dP5taVZJQLIl/HzN6Gv0uDj+gltoqR58+fo/DxqKFc/G6bFx7/2r2hOL79Jom3ciksWYSRKWz9/fTaoOpp6n20OlU179JCbvehVf+H9xzItx1DKT6fn9cdit/WDiHg0b6+K+H32DVh0Nz/HmxQ+GrhRCPCyHeE0KsF0IcIYSoFUIsFUJs1B5rtH2FEOIOIcQmIcRaIcRhOc+zUNt/oxBiYaUuatBJREhixevu1t/HUVWR2QAT6qu4/5p5hGIprlq0gpYO88ZRToKxVGc6b4UZEXCxaOFcWsMJrntw9bBuQzDgaAuzhMWFy95HMdfU0yG2F7augHQSNv8T/v4t+NUsuPNIePGHKsa3cy2seTTvy/lcdu6/Zh6jAm4+e/8q1n0SLPcV7TPFWgC/Ap6TUh4EHAKsB74NLJNSTgGWab8DnAFM0f5dD9wJIISoBW4B5gPzgFt00RhuyERYW2X0ctMYMVN9eMrMwaMDLLp6Ltvbolx936qs39pk3wlGk727DCrEwaMD3H7JbNZs28vX/7wGKc1U0rKgVejLnF5APZh8Elgd8PRX4GeT4MFPwRv3Q+N0OOdXcNMG+K/XYOQh8O9fQSa/QNd7nTz0ufl4nTY+c+9KPmwOl/GC9p2CAiCECADHAosApJQJKeVe4FzgAW23B4DztJ/PBR6UiuVAtRBiJHAasFRK2SqlbAOWAqeX9WqGCKl4pG8/44iZqtgkurfsrztvQi13XnkY63cEzdVjGQkNoAWgc/rBI/jW6QfxzNod3P7CxgF97WGLVqEvHFV97+P0wfRzIdYO08+BS/8E39wMlz8Gh18NvhEqg++oG6H1A3jvmYIvO7razUPXzicjJVcuWsHO9qFTL1CMBTAB2APcJ4R4UwjxByFEFdAkpdyh7bMT0NMWRgNbc47fpm3ra/uwIxXrICodvfsZ9UDwrncr8tonHtTEzy86hOWbW/nSn94klc4UPsgkL8EyxAD6w+ePncjFc8Zwx7KN/PXN7QP++kOSVAJ2r+/fsZoFYHHmsQAAPv0H+PpGOPe3cNBZynXbnennQs0EePWXKj5QgMmNXu6/Zi5t4QSfuXcFeyNDo8izGAGwAYcBd0opDwXCdLp7AJDKRi2LnSqEuF4IsVoIsXrPnj3leMoBJxUPE8XVtwUAFR0Red6ho7n1UzN4Yf0uvv3k26YLYR8JxVJF1wCUEyEEPzxvJvMn1PLNx9fyxkdDM5A4YKRTsPgquPMotUIvFS0GYHXmsQB0ChXjWaxw1Ffgk//Ah/8q6uVnjanmnoVz+LAlwtX3rSIcTxU+qMIUIwDbgG1SyhXa74+jBGGX5tpBe9yt/X07MDbn+DHatr62d0FKebeUco6Uck5DQ0Mp1zJkyMTDygXUWwzA1wRVjRWfEbzwyPF84bhJPP7GNjbu7qjoaw13BjoGkIvDZuGuKw9ndI2b6x98w5DtBsqClPDsTfD+cyDT/RuvqtXn2FxlSpU+5HL1XX719qIPOXJSPb+57FDe2rqX+1/7sDznsQ8UFAAp5U5gqxDiQG3TScA6YAmgZ/IsBJ7Sfl4CfEbLBloAtGuuon8ApwoharTg76natmGHTESUC6ivm0aFAsHduXyeqjVYsbml4q81XImn0sRTmQGPAeRSU+Xg15cdSks4wdJ1uwbtPAaVV36ugrEjZ6vfE/0IpmoV+g53mQTA7oIFX1AzhXesKfqwU2eMYPbYav6htY8YTIrNAvoy8LAQYi0wG/gx8H/AKUKIjcDJ2u8AzwKbgU3APcAXAaSUrcAPgFXav9u0bcMOkVQdB/v0G4+YCXveU/7MCjK21s3IgIvlW4blf/OAoHcCHYwYQC4zRvlp9DlZsXk/fC/ffBhe+iHMuhRO/B+1LV66VZvRRMPZvUBzX5hzLTh8KiOoBE6Z3sTabe2DHhAuSgCklG9pbplZUsrzpJRtUsoWKeVJUsopUsqT9Zu5lv1zg5RykpRyppRydc7z3CulnKz9u69SFzXYiFRMywLqY9U4YiakE9D8fmXPQwjmT6hlxeZWMw7QT3QBGEwLALT3cmIdyze37F/v5aYXVErmxOPhU79WWToAiVDJTxWPqmNcVWUUAHc1zLkG3v0LtG4p+rBTp6ucmaXrB9eiMyuBK4AlFSEqnX33jhmAQLDOvAl1NHfE2TLE8o+Ngt4K2ucc/N7+CybWsjsU58OWyGCfysDwyVuweCE0TIOLHwKbA5ya+6YfFkAsrATA7Slzi40FXwSLDV7/TdGHTG70Mr7OM+guPVMAKoAtHSVpyTMisG4y2NwDIgDzJ9YCsMJ0A/WLol1AH73W//TEIpk/QXV83S9iOm0fwZ8uBncNXPHnzu6cekpmonQBSETVMW5vmQXAPxJmXQJv/hE6istcFEJwyvQmXv+geVCLNk0BqAD2dIy0zd33Dhar6gs0AIHgifVV1Hud+8dNowJ0dgLN4wJ69y9w35nwuwXwyGWwdVVFzmVSg/ZeDncxj7TCHz8NqRhc8bi6weo4dBdQ6RZtMhoiLQX+qgJ1AP3hqK+q3kAr7iq8r5Tw8QpOmxogmZb88/3BS3c3BaDcpBJYSavpX/kYMVNZABX252bjAFvMOEB/CMb0aWB9WAAfvARPXAdj58PxN8PHr8Oik+GBc9Tfyvh/3hnTGcZxgGQUHrlUVctf9ig0HtT171kXUOkxgFQsTAQXfk8F3Hn1U2Da2bDqnvzntud91V7i3lM5/F/XMsojB9UNZApAuUkW0W8ElADE9kL7toqf0vyJtexoj7GtzewUWip5g8Db/wOPXQn1U+HyR+H4b8ON78CpP1Jf9IfOU50j1z8DmfJUZM+fWMsn7TG2tnZ7L6WseFbZgPDC/8LWlXDB3XDAkT3/bnOCxd4vF1AmrrLzKjar+aivqQK1Nx7o+bdEBJbdphrKfbIG5l6H2LqC+9y38+/3tpEcpIp9cx5AudEEIG+/EehsCbHurzD/C2CtXJBR9x0v39zC2NoKmL/DmGA0iRDgdXT7qjRvhIcvBE8tXPmE8lWDWqEe+SWYdx289Sf49y/hsSug4SA4+VY4sJ/tr4I74B/f4cLWT5jl2E3gPgG2pFoxJyLa507CeXfB7Mv26ZoHjXQK1i6Ggy+AGef1vZ/T2880UDUMpq5SAjDmcBh/DLz+W5h3vQpag5oz/PdvKKvmkMvglNvA2wijDuXAp77IT9K/YNUHczhy6sj8z18BTAug3CT0cvMiLIDaSfD89+AX0+Dv34ZP3qyIS2hKoxpCMux9xxUgGEvhddqwWHJaA7Rvh4fOB2GBq/7a1UetY3Oq9MAvvQGfXqTe18eugI9X9Ny3EOkkPH4NbPg7boeVDkuAD8UYGHckHHS2ep1jv64yUfZUNhBdUT76N0RbYXqemz+Aw9svC0DNA3ZS1V3My8lRN0LoE3h7sbrhP3I5PHKJcglf/Tc4/y518wc49AoSp/2cU6z/IfDsF3vOIBgATAug3BTbb8Tuhi8uh01LVV/x1YtgxZ1qpTjrEph1MQTGlOWULBbB3PG1rDQFoGSCsW5tICKt8McLVDfXq5+Bukn5n8Bqg5kXwuST4e7j4M8L4fOvdN4EimHZrSq28OlFiJkX8seH3uCdT9p59YITu+63+t5++caHDOuXqOy4ySfl36+fAmBJRUhYXF3FvNxMPgmaZsILt6r3Qgi14l/wxV6tfMcR17F41UYubv09cskNiHPvBMvArctNC6DMSC07wd59HnBv2Byq2+AlD8HX34ezbwdXtfrC336wCiSWKaNk/sQ6Pm6NsKPdjAOUQpdW0ImwSk1s3QyX/QlGzS7+idzVcMkfVQ+bxz9b/Gpv/TNqTu3czykhQcUBtrVF2dbWrR7A6euXa2RIkMmoa51ycu/dN3PppwvIkoqStOTJzisHQihrLLxbXcsNK1WGUD4X7xFf5v9LXohY86jqdzSAAX5TAMpMXMs1dpRabu6ugTmfhWv/AV95UwUU97wPT3y2LMG9+RO0eoD9sZXAPhCMaq2g00lY/BnYrrl0Jhxb+pONmAln/1J1j1x2a+H9W7fAX78Iow6F036c3dxZD9DtvXT4jGsBbFsFHTth2rmF9+2nBWBNRUlZKywAoOIXN72vBL96bMHdT5zWyG8y57Ny9EJlxf3juwMmAqYAlJmoVm3o8OxDuXntRCUA5/5W+RHffGifz2vaSD8+l82MA5RIKJbC77SqG/GmF5SVNv1T/X/C2Zep/jGv3QHrnup7v2RMCY4QcNEDKqagcdAIHwG3nRVbutV2OA0sAOuXqElcU08rvG8/LQB7JkomX31OOfE1Fd5Ho97r5PBxtfxv+EKVELL8t/DSjyp4cp2YAlBmYhG93LwIF1AhJp8EYxeoQdTJfXPdWLU4QI+bhklegrEk4617VFDvqBvVVKh95fSfwOg58NcblJXXG899SxUKnv97qDmgy58sFsE8rbajC05fv3rkDDpSwrolqt+Pq4gq3X5aAI5MrHB69iBxyvQm1u0MsW3+9+Gwz6jv/Cs/r/jrmgJQZuK6AFSVodxcCDjxexDaAav3vXfe/Am1bN4T3n97yveDUCxFg1XztY87ojxPanPCxQ+oGNBjV/Zcza55VLU+PvprfaaNzp9Qy0ctka7dJJ1eY1oAO9ZA+8cwrUjLqp8C4JQxcAxdAQB4Yf1u5SacebHqGFBhTAEoM4mYCgJXlavfyIRjYMJx8Oov9jnAN3+i8h2b2UDFIaUkFEtSY9Nuss4ydpEMjIEL74WWjbDky50+393r4ZmvwQFHwwnf6/PwBdp72cWiM6oLaP0SEFaVEFEM/XEBSYlLxoesAExs8DKpoUp1B7VYVSHc0V+r+OuaAlBmUloQuMpbxpvFid+D8B5Yefc+Pc3Bo/x4HFZTAIoknEiTkVBtrYAAgHJ5nPR9ePdJWH6nunk/dpVa4V64SKWQ9oEe01meGwg2YhaQ7v4Zf7QqqisGhxcySdV7p0hisSg2kcHqLNMwmApwyvQRrNjcSns0WXgkZZkwBaDMpONhktKK31uGGIDO2Hkw5TQ1dKI/s1A1bFYLhx9QU9ZMoBff28Vf3qx8O4vBQG8E5xcVEgBQcYWDzlYFgQ9fDK0fKMvANyLvYdmYTm6TP4cPUlGVsdQP3t7Wzj2vbB7YPkN73lNW0LRzij9Gfx9KELtgUH1vipoHPEicMr2JVEby8obdhXcuE6YAlJl0IkIUZ/kHiJzwHdU76PXf7dPTLJhYx4ZdIVrD5ekbc9fLm7lp8Rre+KgfM1oHih1r4Z0nSj4sGNX6AAktAO8scxthUCu9834HNePh49eUtTfhmKIOXTCxls3NYXYHuwlUP91Aj6z6mB89u57n3hnAUYXrnwZEaQKQbQld/HWGO4Lq0HKNg6wAh46tpt7r5PkBbA5nCkCZkYkwMZw4bWUO4IyarYJkr/9WVaP2E70eYNWH5bECWsJxMhJuWvwWkcTAl7IXxbJb4W83lXyYbgFUSS0IXAkLAMAVgCsfhzN/rhqKFUm2HkB36WWnZfXPDbQnpFwqtyx5V7khBoJ1S1Qn1QIWTxccpQ+FCYeGvgBYLIKTpzXyzw17iKfSA/OaA/Iq+xEiESEmnIV37A8nfEd9uUucP5rLzDEBnDZL2dxAreEEM0b5+ag1wo/+NgT70CSj8OGrEAuWXFyjt4J2ywjYXJ3NvSpB7UTVQK6ENgAzRvnxOm2dgeB9aJUM0NwRZ3S1m+aOOD977r1+PUdJtG6GXW+XXlehX2cJMwEiYSUAzn2pzxkATpneREc81TW2U0GK+rQJIT4UQrwthHhLCLFa21YrhFgqhNioPdZo24UQ4g4hxCYhxFohxGE5z7NQ23+jEGJhZS5pcBGpKIlKlZs3ToOZF8GK30Oof2ai02blsHE1ZakHSGcke6NJTprWxOeOnsDDKz7mpQH0XxbFh/9Wg0VkuuQhInoraFcmXLnV/z7QI6azjy6g5o44c8fX8Nmj1HtZLiuxT9YtUY+luH8gZyhM8deZrc8p5zzgCnDU5Ho8DitL1w2MG64UC+AEKeVsKeUc7fdvA8uklFOAZdrvAGcAU7R/1wN3ghIM4BZgPjAPuEUXjeGENRUhZXFV7gWO/7YaKP/q7f1+ivkTa1m3I7jPZn5bJIGUUFfl4KZTD2Rqk5dvPb6WtjLFF8rCpqWdP8eDJR2qzwN2piNDUgBAvZcbd3fQ3BHvjFH0MxOoOZSg3uvka6dMZXS1m5uffLuyroj1T8PI2VA9rrTj+jEXWK/P8QxxAXDZrRw7pYEX1u0mk6l8MH5fXEDnAvrkgweA83K2PygVy4FqIcRI4DRgqZSyVUrZBiwF+tkcfQDYukoN/CgRWzpKqpLl5nWTYPblqntoP4fJzJ9Qh5Twxkf7tsLTb/Q1VQ5cdiu/uHg2bZEE33vqnaEzsWrTCyrHHEpeGQc1C8Ce6hi6AjAhp7Yj6xsvTegAwvEU0WSaep+TKqeNH553MJt2d/D7f24u5+l20r4dtq/uX1uNfswFTkSV9ecp9zzgCnDK9CZ2BmO8vb3/GX/FUqwASOB5IcQbQojrtW1NUsod2s87Ab35xWhga86x27RtfW3vghDieiHEaiHE6j17BmlWZrgZHv40LP1+yYfaMvHK9xs57pvKn93PUvFDx1XjsFp4Y9MnqvCon9OqWjQBqKtSvvGDRwe48eSp/G3tDpas+aRfz1lWWrdAyyaVbw8qDlACwVgSh82CNdFRmQygMjBrTAC33arSQffBBaQHgBu8Kn51wkGNnHPIKH7z4iY27a5AbcH6p9VjMc3fuuMoPQ00FVP72l1DNwisc+JBjVgtYkBGRRYrAEdLKQ9DuXduEEJ0aYUo1XKvLEs+KeXdUso5Uso5DQ0N5XjK0nnxhyrfPlp6aqMjE0VWWgCqx6meNG8+pG5yxRJtgw3P4Xr5Vp723MqNq09Wg8w/eLFfp6GnktZWdQZHP3/sRA4dV83//PWdwW89vekF9Xjwp9VjvLQVVSiWwu+yqRX1ELUA7FYLc8bXqEygfcgCau5QAlDv60xg+P7Z03HZLXznybfL745Y/zQ0Tof6yaUfmw0ClyAAcS3+Yx+6dQA6NVUOjp/aMCBjIotKVpdSbtcedwsh/oLy4e8SQoyUUu7QXDx69G87kNsDdYy2bTtwfLftL+/T2fdBezTJ95/kugH8AAAgAElEQVR6h++cOY0mf4n++B1rVB8WYSm56EpKqfUbGYAP2TE3KQF48joYt0AN0rDn/NN/T8Xg4+VqoMjudepYix1v1TQWx47nCusLENzer1PobgGACkzefvFszvjVv/jm42t54Jp5lR3AkY9Ny1R+/ahD1e+lWgDRJD6XXROAoWkBgErt/fnz79OWclAD/bIAsgLg7XwvG3xOvnvWNL71xNssXr2VS+eV6Kvvi449qubh2G/07/h+zAXO6AIwRFtBdOcPC+cgBqAauKAACCGqAIuUMqT9fCpwG7AEWAj8n/ao97ZdAnxJCPEoKuDbronEP4Af5wR+TwVuLuvVaGxpDvPCul28tXUvD39uPmNqinzTpYRnvwmeOphyKrz3t5JeN5bM4CaBGIiOg/6RcPL/wj9/BjvfURWgfeHwqmriGRfAAUfA6MPZ/GEHP1n0shKAfviMAVo7OmMAuYyvr+K7Z03je399h4eWf8TCI8f36/n3iVQctryi2i/rHSZLvDFmLYBwaMhaANDZ42nFR0FOt3v65wLS3kvdBaRz8ZyxPPmf7fz42fWcOK2RRl//ExwyGcmTb25n5MZHOEpmiE45i37byqX2A9IzwIZoN9DuDMTNH4qzAJqAv2gnZAP+JKV8TgixClgshLgW+Ai4WNv/WeBMYBMQAa4BkFK2CiF+AOgjrm6TUlYkz2z22Goe+tx8rr53JRff9ToPX7eACfVFrMrfeQK2Lodz7lAB1nhQ+ceLzM0OxpLUEsdSaB5wuVjwX+ofKPFKxVTeu/6YjCpLpn5qj74yh42zE7O4kQhEP9tLtIbj+F027Nae/z9XzB/H0nW7+Mnf13P0lHomNQyw7/Xj1yEZhsmn5PjGSxO6UCyJz2mD1qEtALP02o4tLZzez4ZwzaE4QnR154G6Ef34gpmc8ct/cdvT6/jN5Yf18Qz5aemIc9Of1/Dyhj08aH+KLaKJk367lcmNe5k5upqZo/3MHFPN9JF+3A4rUkriqQzBWJJgNKU9JpUou+0cV2pH0GSEpLBjH4AOm0aioABIKTcDh/SyvQXoMbxTiwfc0Mdz3QvcW/ppls5h42p45PoFXLVoJRfd9ToPf24+B47I8yVOhOH5/1FpaYdeqZpzIVWusStQ1GsGO8I0iTS2wQg0CdHp/imCKqeNg0fXEGn2UFWia0SnNZLsccPoPB3Bzy6cxWm/fIWr71vJrDHVeB02vC4bVU4bXqdVe7RR43Fw+AE1VDnL2D5j41I1YGTCMcodhuhHEDjFWJ+ATGpIC4DTZu2sB3D0ryV0c0ecWo8DWy9iPqnBy5dOnMwvlr7PBYft4sSDih92AvD6By3c+NibtEWS/PTMsRzz8no+mnoNX66bytvb2/nn+7t54j8qo81qEVS77YRiKRJ5fOCbRldhK+E6LckICYubPIMZ90uG9VD4GaMCLP78Ai6/ZwWX3v06D352PjPH9HEz/9cvIPQJXHS/asequw1iwaIFINyhPpCDIgD9YP7EWtp2u3FF99KfdVFrON6nAAA0+V386tJDuX3p+6zfESQcT9ERSxFO9Mwtd9gsHD25nlOnN3HStCYafPmrqSOJFKs+bOPfm5pZvrmFw8bVcMs50ztN503L4IAjO+MxTl+/LIB6W6bz+CHM/Al1/HLZ+6TGebH1Iwi8JxSn3tv3//kXjpvE02s+4Xt/eYcfXyA4ZkoD1gKxnXRG8usXN3LHso2Mr6vi3qvnMmP3s5BJMf7oy/jamKmAip3tDMZYu62dt7e10xpJ4HfZ8btt+F12fC4bfrcdv8vO2m17ufXpdSSsHmwlFPZZ01FSjgGaBmYghrUAAExu9PHnLxzB5fes4PJ7lnPfNXOZM75b29nWzWpE36xLYNx8tS1bVFP8TSPbcKqYgfBDgAUT6gi97sHT1kKRjXi70NKR4BLbP+H1t+CIXo0+jpvawHFTu2ZzZTKScCJFOJ6mI55iR3uUF9/bzdJ1u3jxvd0I8TaHjavh1OlNnDpjBBPqq0imM6zZupd/b2rh3x808+bHbSTTErtVMKnBy/2vfYjTbuHmM6Yp992e9XDoFZ0v6vSXXgcQTVHv0PobFbkIGCzmT6xFvgBB6aa2nxZAva9vMXfYLPy/iw7h2vtXcfV9qxgVcHHh4WO4aM5Yxtb2dHnuCsb46qNvsnxzK+cfOpofnHcw3tgu+M+D4B8DoztdSUIIRgbcjAy4OW1G/p5Aur7HhBtPkUKXSmewZWKkB2ocpIEY9gIAcEBdFX/+whFc8YcVXLVoJX9YOIejJtd37vCP76msgpNzBnXrX/gS3AZRrdpwqPcb0ZnS5GU7Huoie/t1fGs4wXH2pbBsAxx6VXHj/FBNr3wuu8qwASY3ejlmSgPfP3s663eEeH7dTi1+8B4/+ft7jK/zsCcUJ5xIIwQcPCrAZ4+ewFGT6pk7vhaX3cL3n3qX3/9zM7UeB5/3/ku90OSTO1/U6SspqyuZzhBNpqmxpjqPH8LMHluNRUBr0kGtKD201tyR4NBx1QVf4/WbT+KF9bt4bNVWfv3SJu54cRNHT67n4rljOXV6Ey67lZc37Oa/F68hmkhz+7kTOc+5GvHI91RPJiScclu/+93rWX1h3NTGi7vOUCyFhzgZmzECwAPJfiEAAKOq3Tz2+QVc9YeVXHP/Ku668jDly9y0DDb8DU66RWXW6GRdQMXfNGLaQHiXQQSgwedkvfQg+pEFJKWkLZLAGwiroPO6p+Cwq/bpfIQQTB/lZ/ooPzeePJVtbRFeWLeLf21s5ugp9Rw1qZ4jJtVR7em5Ur31UzNoiyT4yd/f46zxSxjjHwMNB3Xu4PKXZM3pfYAqNgymzLjsVhp8Ttozrn7HAPK5gHQcNgtnzhzJmTNHsn1vlMdXb2Px6q185ZE3CbjtzB1fy8vrt3N57Ua+Mf4tfMuWqs9H7UQ4/maYdZH6uZ/oWUoh6So6CNweTeIRceRApGcbjP1GAAAafS4evX4Bn7l3JZ9/6A2+e/pkFq65GVEzoacLw6lZACXcNOLaNDB3lTFiAE6blZjViy25o/DO3QjGUiTTEk9a+xKueXSfBaA7Y2o8XH3UBK4+akLBfS0WwS8unk1HJEpg62tsm3gOY3JXmU6/mqpWJHor6EAlh8GUmREBN21RJ6RLE4BwPEUkkS5KAHIZXe3mqydP4csnTua1D1pYvOojZm64g19VvUxVZC98Uqssw0MuhdGHl2XKlcNmobbKQXvaUbTQtUeTuIkj7INUWDqE2e/aQddUOXj4uvkcMamej5+7A9G8gXWH3KyKS3LJuoBKcBtoAuBwD/2bhU7a7sOZKn3FqPcBcqZDqlXyR6/C3o/LfXol4bBZuPO4ND4R5SebxrA8d1qWq7QYQHYYjEUfBjP039MRfifNieJvjDp6EVihwHtfWCyCo6fUc8eCDq4Tf6Vq4ny47FH4+vtw1s9hzJyyjjhs9DlpSzmL7u4ajCXxDGR6toHY7wQAwO+y88DF47nZ/ReWWw7lzOeq+PxDq/m4JdK5Uz9cQMmYsYpNAKTTjzMTKbkfUEs4gZU09lQYDr5QbVz7WAXOsDTcH72ItNjYWj2Pzz2wmnf0hlolZgH1HAYzdCuBdUYG3OyK20uel9tbFXC/eOcJlYZ68QNw4BlgrUzSZaPfRUuy+OtsjyZxi/iQngc8WOyXAkAmjXj+f7BnYhx63Z1847SD+NfGZk7+xT/52XPvEY6nlEVgdZZ000jrlYkGEgDhDmAlU3L/mNZwAh/azXHETDjgKFjzWMlDV8rOphcQYxfw+88dT8Bt5+r7VrKlOaxu4CUE9PVhMJ5KTwMrI01+Fy0pbRVfghWwJ6SsuVJdQF1IJ2H9EnXjL7IWpb80+ZxK6KCoauD2aBIPMUM0ghto9i8BkBI2/B3uOhrWPgpHfAnnyGnccMJkXrzpeM6eNZLfvfwBJ/z8ZZ78zzakq7SbhtH6jQDYPFrmR4mB4NZwHL/Qbo6ugEqhbdnYrxbaZSO0E3a+DZNPYmTAzYPXziMj4apFK+gQHkjHi14Z662g3TKiFgLdXYRDkJEBF2Gp3XxLeD/31QUEwOZ/qmaDMy7o/3MUSZPfxW5dAIoYChOMqiwg+xAeBzlY7D8C8NFrcO9p8Mil6iZw0f0q80djRMDFLy6ZzZNfPJKR1W7+e/EaWlKuklxAmYTmL67wCqicOLyqNVMkVFrn05ZwAj+a4LmrYcZ56ka59tFyn2LxbFqmHqecAqgK1vuvmUtbOMGja7X3sciVcecwmKE5Daw3mvwuOvTuOiX0ydEFIF9RX0HefVIlTkzu0Ryg7DT5nYQyuqVThAUQieMSSWwGqc8ZSIa/AOx8Bx6+GO47QwUpz/4l3LACZpzfa4+fw8bV8Jf/OpKjJ9fTnHKVtJISSeO0nNXx+JQAtLeWNnuhtSNBg03LkHEF1L+DzoS3H4fUIE0E27QUvCOg6eDsplljqrnyiAPYoOtbkYIeyg6DMY4AjAy46EBr1laSC0hVdPfW06koUnFY/wxMO3tALKUGn4uwfp1FBIKjWnq2MNNAezB8BaB1Czx5vXL3bF2uOmd++T8w55qCwSmLRTCxoYq2tKskF5BIRkljrezw8DJTFVCdJEPtpRUPtYYTjHJp7hQ9Y+qQyyDa2tmHfyBJp+CDl1TxV7eMk5F+F3szpblGQrEUXqcNS2JoN4LLZUTARUfWBVS8AKgagH34zG5apmYtDID7B5QFkL3OIlxA8ai2j4FicwPF8KwD+PBVePA8sNjg6BvhqK+Cu7TxwyMCLlrTbjKx9qJUUkqJJRUl6XD1q6/OYBGoUQIQLlUAIgkOd8QhSacATDoRqhpgzSPKGhhItr8Bsb29uiCa/C5CaF/+IgU9GEuqVtCxoT0LIBeX3Ypw+dVophKC+s0diX0LAL/7JLhrYeJx/X+OEmjy51gARbiAElp69oDM6TAYw9MCGDNXFXZ95U218i/x5g/KnA5JD5locS6DeCqDU8ZIW43j/weorlEtMWIdpVsADXbdBaQFkq12lRL6/nP9mqa2T2x6QbW+nnRCjz81+tV7CRS9Mg7F9GEwoaJbXAwFqnylB/WLrQLulUQE3ntWzfatUNpnd+q9TsJ6rKMIodPrc4wUmxsohqcA2Jxwyq1dWzuUiL5qLLZNQlDLNU4brN+ILgDJcGn9gFo6EtRZo2rgeu7K6pBLIZ2Ad/9SztMszKalSvh7EftGn5MQpbmAgtEUfvfQHgfZG76Adv2luIAKdALNy8bn1dyFAXL/gCr4c7qLnwucjukCYFoA3RmeAlAGRgbcBKUHayqi/MsF0KsNKz4PuMxYHG4S2EhHSxOA1nCCGktYuX9yfe4jD1E9eNYMYDZQxx745M2uzd9yaPQ7Cek+4yJdQKF4jgVgIAGoCVSTQRSdBRRJqPbc/U4BffdJqGqE8Uf37/h+UuXXhK4ICyBtwPTsgcIUgD4Ykes3LmLV2B5N4SZuyA9ZWHhL63qaSBNNplUdQPc2yUIoK2DrCtVmeyDY/JJ67EMAnDYrNndpvZ2C0RR+p9VwAtBU7SEsXaSLfD+bs0Vg/QgCx0Pw/vMqBXiAJ23V+L0ksRUUACkl0mDjIAcSUwD6wO2wknZoX/wiUgeDsSRukTBkqlncWoU1UbwAtIRV9o9Xhnvvkz/zYkCoyuBKIiU0b4S3HgZPvZrm1gd1AS9xUXxldyiWpMYpVbsBAwnACK0WINpRnEW3R28D0R8LYMNzahb1ALp/dJr8TiK4Clo64UQap9Sy1Qz43aw0RQuAEMIqhHhTCPGM9vsEIcQKIcQmIcRjQgiHtt2p/b5J+/v4nOe4Wdu+QQhxWrkvptyUsmoMRrWGUwa0AJJ2P7Zk8VkjbWGtTUImrIrAuhMYDROOVUVh5W4NEWuHdUvg6a/CL2fBb+bA5pfhsM/knd3c6HcRwV2UpSOlJBhLUafXORgkCwg6U0ET4eKSF7JVwP2JAbz7JPhGwdj5pR+7jzT5XYSki0yBWIfeBwgwLYBeKCUN9KvAekD/NvwUuF1K+agQ4i7gWuBO7bFNSjlZCHGptt8lQojpwKXADGAU8IIQYqqUsud8wCGC01erxtoXZQGkcBMzZLVh2uHDFW4jnZEFx/xBpwXgTIfANbb3nQ65DP76BeUKGregfyeWyagWzm1bYMu/4INlsHUlyLRqOjbhODj6qzDpJKjN3zK6yeckKD3UFCHm0WSadEZSZ9cK2gwmAGFcBKJFuoCyjeCcypratlq58Ap174zuVZlXc6/LK7yVotHvIixdJCJBPSG0V9ojamEGmBZALxQlAEKIMcBZwI+A/xZq8OqJwOXaLg8A/4sSgHO1nwEeB36j7X8u8KiUMg5sEUJsAuYBr5flSiqAx18Duyhq1aiygBKGmQfcBVcAH9tpixSXD96qtYK2J/LMS552Dvztv1UwOJ8ApJOwbRW0fQh7t0L7x9rjNvUvndO7Z+QhqqZj8skwdl5JaYdNWjHYuFiQQhKnt4I2yjCYXEb63ayVbmSJMYA6rwOe+z2sugc6dqn6mXy89zeV7XXwp/f1lPtFo89JGBfJaCivAOjJGYCZBtoLxVoAvwS+CejfhDpgr5RST4/ZBozWfh4NbAWQUqaEEO3a/qOB5TnPmXvMkMQbUJNyk5G9FLrV6GPnbE7jrTKs7gBuES44GFxHFwBrPgFwepUIvPsknP5/YM/5mkqpmsatfVS1EI7k9O33NkFgLIycBQedBdXj1O+jDwdv/wd6NPmdBKWbVKS9iPdSGwZjMZ4A+N02ohYPItFc1P57OmLUeOyqDYT+Prxwiyroy52p3J13n4TqA7rM9h1ImvwugtJNJlacC0giELZ8UrF/UlAAhBBnA7ullG8IIY6v9AkJIa4HrgcYN25cpV8uLzU16obT0d5KoVKyYEyNnTOin9FeFcBHlA2hONOKKJ1oCSeosqYQqWj+YemzLlEzAt5/TmWKtH0IaxerbS2bVPO4A8+AmRdC43Twj+4qFGWkUQuOpqOFxVzvBOoTxhkGoyOEIGP3Ykt9WNT+zaEcqy/aqqwsdw0s+TJU1cPUXkJ1kVYVdzniS2Ud9FIKTX4nO3Ah4/kLDtu12Jy0exCDdK5DmWIsgKOATwkhzgRcqBjAr4BqIYRNswLGANu1/bcDY4FtQggbEABacrbr5B6TRUp5N3A3wJw5cwa1uXxtnSqSigQLC0BHJIqNtCHTQJ3eWjwiTnN7B1B4ld3akWCMOwUpOquAe2Pi8ao5279+DsvvVD2ZAMYfo1w508/NLyBlpNHnZKP0QLzw1DJ9FoARBQBAOr04I8VNy+pSBRxpBd9IuHAR3H8WLF4IC5+GsXO7HrR+CWRScPDAZ//o6NXAluTWvPsFtVkApv+/dwpGb6SUN0spx0gpx6OCuC9KKa8AXgK0UVAsBJ7Sfl6i/Y729xellFLbfqmWJTQBmAKsLNuVVICRNV7C0kmso3Bbg2y/EQNaAHpH0ODelgJ7KlojCca4tQBpPgGwWGH25apHf7QNTvo+3PgOXP2MytoZoJs/dFZ2W4toHqZ3Aq2S2k3UQEFgAKsrgEtGi8rAau6Id6aARtvAU6sE74rHwTcC/nQR7NnQ9aB3noC6yTBiVgXOvjjsVgsZmwdbKr/QBbWB8MKAC7OBYF/C999CBYQ3oXz8i7Tti4A6bft/A98GkFK+CywG1gHPATcM5Qwg6LxpFNMmIRExbsdBp1fFOjqKbAjXayfQvjj+ZtWF9YYVcMxNUN1H1lCFadDaQdjTEcjk/9jpswDcGX3gjbEEwO7xYyXTOaAoD6oRnFYEFmlVTd0AvI1w1ZNgscNDF0C7Zqx37FbNFmdcMGjunyxOL450NO8u7dEkPmsSYbaB6JWSBEBK+bKU8mzt581SynlSyslSyou07B6klDHt98na3zfnHP8jKeUkKeWBUsq/l/dSyo/fZaODqqIawiWz5eYG/KBpN7hIqHgBGOHImQWQD5sD6iYN+s3CbrUg9cK+ArnjugXgzETA6jDENLBc3F5llbW15X8/o4k0HfGUagORiquePp4cZ2ftRLjycZUG/cdPKwth3VMgM4Pq/tGxuHzYSOWd8haMpfBZEoZ0zQ4EZiVwHoQQxK1VRTWES2cHwhsw1Uy7iceLcHUBtHTEczqBDpwbZ18RruIK+4KxJHarUMVxBvP/A7i1jqDNrfldel1qAPTurd2b6Y08BC59GFo/gEcuU2m9jdOhcVrZz7tU7G7NMstTDdweTeK1JAxpmQ8EpgAUIGX3YUsW9htnzW0jftA0H3eqCFdXMp1RVbJ6jnxvlcBDFLtHE4ACOfJ6K2hhsD5AOn6tUdretvwCsCe3CjiiWQu6CyiXicfB+b+Hj5fD9tWD0vqhN5xV6v1M5Sl6a48aNztvIDAFoABpp19VvOYhlkxjy2g3REO6gNQXKVNExXObVgNQY412OdYIOL3FtUoORlP4XMZrBa0TqFE38UJB/eZQrgWgCYCnFwEA5fI58/+pOo1ZF5XtXPcFj1ctXNrzxK70Fi2mC6h3TAEogMUdwJMJk870nVGhqg01ATCkC0h9kezJDmLJ/AHSFk0AAoQ1/7hxims82sq4UOvrUCyJP9sK2jgCpxOo1oL6ofzX2dyhdQL1OfJbADrzroObNkDN+HKc5j6jD79p29u367I9msQlY6YF0AemABTA5qnGRzTrL+2NoN4KGow5dEJzAflFOO91QqcFUCU7es4CGOL4tMruQvOPgzFjWwDWbFC/kACo97quqggLQGcIvd9+bfhNsL3v62yPJnFIsw6gL0wBKIDTW4NTJNnV0veHTG8FDRjT1LRYSdmq8BFlTyi/AOgWgKevVtBDmEC1mn9cKN21qwVgPAHQzzkZye/S2xOKU+2x47BZcoLABQRgCFGjubo6gr1bALFkmngqjSNtWgB9YQpAAar86kPW3NJ3b5WsnxGM6QICpNOPnzC7CwiA3gfIlQrlLwIbgtTold2h/NlOnTEAowtA/mB3jypgm8tQC5hqzdUV7ehd6IKxJHbSWAxaoT8QmAJQAN1tsLc1jwDEDO4CAtURVBRnAQhRoBPoEKWxppqUtJAokO0UiiXxuw1sAdhcZLAi4yFknmpgJQBaEVi0tdd5ykMZm0u9N4k+hE65ZvXYnEG/lxXGFIAC6BZAsL3vjAp9ILy02FThkwGxegL4tY6g+WgNx6l22xHxdsMJQJ3XSQgPqTyukVQ6QziRptqRVu2OjSgAQpCwVeFIRwjF+55nraqAdQugzVDuH0B1nCWnCr8b7bmWuWkB9IopAAWwaHnukWDffmO9FbRR3T8AFlc1NZZYNje8L1rDCWqrHKo61GACYLNaiAhP3sruDu2GWWfTex0Z6xp1MnYvPhFlV3usz32ac9t/R1sLB4CHGjYnKWx9zj/W+wABZgygD0wBKISWURHL4zcOxvRqQwObmS4/AUthF1BrOEGdx5gCABCzehF5GsLpbSCMOAymC04vXqLs6EMAYsk0Ib0NBGh9gIzlAgKIWz19VgIHY8kc16wpAL1hCkAhtBTJZCRPFlA0ic+aQBjYAsDpx0cxLqAEjR6p3CMGqgLWSdm9eecft2uN4IwuAFZ3gCqi7Az2LgD6+9xgZAsASNmqsKbCpNKZHn8zXUCFMQWgENoqV0bb+wyoqYZTSWPnGrtUwduePm4YOqoTqHHdIxmHD0eqbwHQLQC/QWcB6Njcfrwiys4+LIBsHyCfQ7WNjhowBgBIu4cqYtmitlzaI7kuIAN/NyuIKQCFcHiRCNwyzN5IstddgtEkXovB+424/FhJ0xEO9il0mYykLZJkZLGdQIciTj+eTJhEqueKEYw/DEbH6vIRsMT6tACyVcBep8p2yqQM6QLC6cNDjF29XGcwliRg1b6zpgXQK6YAFMJiIWX34ifSpz9VjYNMGDoIrN/Mnalwdih6d9qjSdIZachOoDpWTwCviPYZ7O4cBqPNAjDYMJgsTh8+EStsARTTB2gIY3X58Ipor/Ur7dEktQ5NAIy8OKsgpgAUQcbpxyei7Az2Pnyis+GUgc3MnHYQezp6v2noVcB1Nr0RnPFWjA5PAB8Rdrf3/V4CeDJGFwA/Xvp2AekxgDpvkX2Ahih2t5+qPiyA9miSGru2mDHyd7OCmAJQBBZXAD9hdrb3vmoMxlK4MLoLSAV0/UT6rAZui2idQC3G6wSq4/LWYBMZmtt6D+p3DoPRx0Ea0wWE04tLRtnVHun1z80dcQJuO06b1dAWgMPjxyti7O7NBRRNUW03LYB8mAJQBFaPqpLdmWfV6JTGrgPQ013zVQO3aH5jP9rN0YACUKVXdrf1XtkdjCXxOKxYkx1qHKLBpoFl0YQrHgn22uG1SxVwxHh9gHQsTq8SgD5cQNkYgCkAvWIKQBFYXAFqrL3nVKuGUxkcmaixzUztZu7Pkwqq9wHyZQXAeO6Rzsru3gv7ejSCG0LdL0tCEwAvUXYHe76fzaGcKmC9EZwBLQCc3j6DwO3RJH5rQrUtt9oG4eSGPgUFQAjhEkKsFEKsEUK8K4S4Vds+QQixQgixSQjxmBDCoW13ar9v0v4+Pue5bta2bxBCnFapiyo7rgDVovecat1lYMsYvOOg5uuutvZdDdwaVtvdmQ6wuQ25OrZoQhdu772wLxQzeCM4HYdqk+Dt43Pb3BGn3pdTAwCGa+4HgMOLnRQt7T1TezsLNA38vawwxVgAceBEKeUhwGzgdCHEAuCnwO1SyslAG3Cttv+1QJu2/XZtP4QQ04FLgRnA6cDvhBDWcl5MxXD68RLuNaAWjCWxkcIqU8b+oGmr+ZHOeN8uoHACr9OGLRE0ZBEYkL2pRzt6twCCeiO4WNC4AWDInruXGDt6cV3u6Yh3FoFFWtXgGyOukrX3M9xt9kE6I1WLFmHw5IwKU1AApEKXV7v2TwInAo9r2x8AztN+Plf7He3vJwkhhLb9USllXEq5BdgEzCvLVYvFfzIAAB6YSURBVFQalx9PJtJrDCA4XKoN7R6w2Gi09y0ARu4DlEUTumS4935AXSwAA7q4sjg7LYDu7pFYMk0oltMGItoKHuNldAFZSycRaSeZUw0c0uo5VHq2gb+XFaaoGIAQwiqEeAvYDSwFPgD2Sin1hPFtwGjt59HAVgDt7+1AXe72Xo7Jfa3rhRCrhRCr9+zZU/oVVQJXAAtp0vFwtlmYjsoA0qoQjfxBEwKcfursscICEN1rXAHQVsapPhrCBaN6DMCY08CyaOdeZ4v3iF111gDoQeBWQwaAgezqXlUDd35u9VoWtzTnAeejKAGQUqallLOBMahV+0GVOiEp5d1SyjlSyjkNDQ2VepnS0G4aPiI93ECq46Dec9zgHzRXgJo8DeFawwnqhokFYE2Ees2OGTYxAO3cR3lSPSyALlXAYNg+QEDW0lGB4M7Prd7TySmjZhuIPJSUBSSl3Au8BBwBVAshdKfhGGC79vN2YCyA9vcA0JK7vZdjhjbZFMmeApBtBQ3GX2m4/PhFhNZIoos5rdMaTlBjdAHQXAb+XtJdpZSdMQCjC4BDnfsIZ7KnBRDKqQIGg1sAWraTiHapBdAFwJGJGTs9u8IUkwXUIISo1n52A6cA61FCcKG220LgKe3nJdrvaH9/UarmMkuAS7UsoQnAFGBluS6komRTJCM9AmrBWHJ4uIAAXAG8MoKUnSmfOlJKWrpYAAYNAmfnH0fYHep6Y4ynMiTTcphYAEroGhyJHjMBdFdJZwxgr+EtgCpi7MoRdL2nkz0TM/7CrIIUE/YfCTygZexYgMVSymeEEOuAR4UQPwTeBBZp+y8CHhJCbAJaUZk/SCnfFUIsBtYBKeAGKWVPG3wo4tQEQER6mNPBaBK/dZj0HHf6cWd2AqpVQJPflf1TOJEmkcpQ67Eb2wJAzT/2xqJdXAbQ2QYi4JCQjhtbAGxOsDqpsyXYFYqTzkisFlXT0KUNRDoF8XYDWwBKAHx9WADWVMR0AeWhoABIKdcCh/ayfTO9ZPFIKWPARX0814+AH5V+moOM5gIa4Uz0MKeDsSR1jjRkMP5KwxXAmVbDUrq7R1o1v3GjKwUybWgBEO4AvmCkR358UKvpqM3OAjDuNQLg9FJti5HOSFo64jRqgt7cEcfvsqk2EB1aooURO4FCVgAanakuBW+dAhA1/veygpiVwMWgBYFHuZO9BIFTnQ2njL7ScAWwJZUAdHePtGhFYI0G7gSqY3X58YteLADNbVBtMfYwmCxOX3auQe7Cpbkj0bMIzOAuoEZnil05n9lgNKksnmTE+JZ5BTEFoBi0m91IZ8+UumAsmdNwyuDBJqcfS6IDC5keFoDeCK7OZnwBEC4/tdaeDcT0qu6AwaeBZXH6qEJdS661s6cj3jUADMa1AGxOsNipcyR6ZAFVu6yIlMFbtFQYUwCKwe4Gi40Ge7zXGMCwGTqhC50r0UMA9EZwNULrLmnUSmAApx+/JdplxQidMQCjD4PJ4vDh0tpa7+xiAcS7joIE41oAAE4vtbYEe0JdYwCNbm2wkWkB9IkpAMWgFUnVWmO0hBNd8seDsRR+m24BGHylocU6xlelevQD0rOC/BbjdgLN4vLjJdLDBaRbAN7sMBiDC4DThz3Vgc0iuloAoXhOBpBxO4FmcfgIWOM0d3SmLwdjKRqc5iyAQpgCUCyuAAFt9ZsbbApGk/gsCRBWsNoH6+zKg3ZTH+NJ9gwChxM4bBZcqVCXfQ2J04cnE+7FBaSE3K0LgJGvEcDpQyQ6aPK7shaA3gaiSxUwGNsCcFTh04ox9c9tezRJvS4ARnfNVhBTAIpFWzVCV39qtuOgo8q4rYN19GC3q6cA6DUAIhZUG4xaBwDgDGCXCaKxGNFErjWnAofOtMGHweg4vRAPMSLQKQD6VLcuVcAWezabxpBoLaGB7FyAYFTLzgPTBZQHUwCKxenHI9WNQS8Gi6fSxJIZqoZLwyltxTvC0bMhXJdGcGDsTpkuvVNm12IwvQ2EiGtWjuEFwKcEwO/KLlp6rwKuMfbixeHFLdV3Uo/RBaNJavTkDNMF1CemABSLK4AjpZqi6qsp3WfsFgafBqaj3RgbHHHCiTThnMZ3XQTA4TNm62AdvbdTt1TQzkZwIbDYwObq6xmMgdMPqRijfDZ2tseQUnY2gvMNgz5AOk4vDs1q2x1U19llHvBwWJxVCFMAisUVwBoP4nXasqmgetaIW8aGxypDK3yq1Wb+5loBnQJg4E6gOtrK3kfXyu4ejeCMvCqGrFtnbFWaaDJNMJrq2QYi0mbsADCAw4c1FcYiYFcwTiSRJpWRVNu0diZGz86rIKYAFIvTD/EgIwKuTjNTHyAuDT4QXkezAKp1AejoTQCM3QYCyF5n99YewVhSEwCDt4LW0TuCutXndGews9V3XZUWBB4OFoCjChHvoMHnZHcoli3o81mHSYFmBTEFoFhcfoiHGOV39LAAHHKYdBy02sHuwSeUOa3fLOKpNB3xlPFbQetoLqAaa7zLMPFQLJUzD9jg1whZARjpVp/THe1RmjsS+Fw2XHZtGF+0zbhFYDpOLyTCNPld7ArGs20gfBbTAiiEKQDF4goAknHedI8YgC09jKoNXQGqtDRIXQD0GoDaKufwcAFpFsBod7KrBRBN4nMNg1bQOlqbhHq7ev92BWNdR0FKqYLAhrcAvJBJMsor2BWM0R5RAuC16E0ah8l3swKYAlAs2qpxnEcVSaXSmaypaUsPEwsAwOnHmerAahFZAdCrgLMuICNXAUP2vRzhjHep6QjFUvjdw8kFpK6z1qaucUd7jOZQThuIZER1PTV6DEB7r8ZUqRYmumvWI3QBGCbfzQpgCkCx5Kwa0xlJc0ci6wKypIZRwylXABFvp97ryApAtg+Q1wHR4eMCanQksu0g0hlJKJ4aZhaAugZ7Kky918GuoBqb2BkANngfIB0t2D3anaYlnMgGut3EAWEKQB5MASiWnBx5UP7UYCyJzSIQycgwcgH5IRakwefMBoGzLiCPtjo2ugDYHGBzUWfrtAD0Wc/+4TAMRkcv7tKKwXa0qyBwtgp4OPQBgqyra4RLLcg+2K3StV0yphZmRs/mqiCmABSLtmqsd2jdFdtjBKMp/G47IhkdPqsMpx9i7TR4ndkiKd0FVGeNA9L4AgDg9FNtjdIRT9ERT2XbQPhddogNFxeQdg1aMdjHrRGCsVQvnUANLgDa4qvBpUR8oyYADnMaWEFMASgW7aZXp03/2hlU6WY1TiCTHD6BJlcA4poFkBMEtloEfjEMGsHp5PTK3/3/t3e2sW1e1x3/HVEiKUuiJNvySxynTTYHXVBsruGlyZquQYrl7YszYMgWYK1XZPCKpVuK9cPSfkmWIlgwbO1QYEiRoV5doGsWLNkSDAEyIwvQ9UOzOFma125212Sx48jOZOtdJCWdfbj3PnxI640KJZKH9wcIJC8p6T4Pyed/z8s9Z8KJOcBgdtH5xfNtvNM5EASgNMWuwTw//8C9f9trC8G1uwXg+wJv90UZT45OMpDrpiv2AliVKABrxV/0+nSabHeXtwDK7Mj7eiNWVhp5bwEM5PhgqsTiousFPLylh66iLwPRznWAAvkCfb60x+hEMbEAhkJ7z3YudRHoyrgLoLcA1FdH3l5bCrrdLQDvAhr22U7vjc9R6O1xQW4rrtkNYi1N4feKyPMi8qaIvCEi9/rxrSJyXERO+tthPy4i8i0ROSUir4rIgdTfOuxff1JEDi/3P1sSf0GQ4ji7Cs6fOjE3z9ZcKDhlxAWUH4SFErt6XWD0wkyJselidR0gExZAIamV7zYPeQvASjewQG7Ab2CsfD4rlUBDKWgbQeABmU36HicCEC2AFVmLBTAPfEVVrwGuA+4RkWuA+4DnVHUf8Jx/DHAbsM//HAEeAScYwP3AJ3G9hO8PotEW+MAhcxNJdcWJ2TLbrLSDDIQUSR9QOz9V5MJ02Z4A5AtJbafRiblKDMBKM5hAbgCKU+wqVOoajaTrAGUH3Ge7nfHvVVdpOtnjMNjbDaUZO5b5BrGqAKjqWVV92d+fBN4C9gCHgGP+ZceAO/z9Q8D31PFjYEhEdgO3AMdVdUxVLwDHgVsbejQbjS8HsXswn8QAtvZYcwE5986uXKW++v9NF9nWl7MlALkCXaVJtmQznJsoJim9fRgTgGylJHTgkkqg7U5w85Qm2Vlwx1bI90B5OloAq1BXDEBEPgp8AngB2KmqZ/1T7wM7/f09wLupXzvtx5Ybbx/ygzA3nlgA47Nlhrzf0Y4LyFkA2/zmofOTRcamSwz39VQEoN03ggHkCsjchCsfMFlMdnWHkt9mBCA3kASBAQZy6TIQY7DFgAD4vsCUphkZcMc52NvjLIAoACuyZgEQkX7gCeDLqjqRfk5VFdBGTEhEjojICRE5cf78+Ub8ycbhc+R3FfKUFhaZKy8yZKUdZCBfXRH07PgcF2fLrgzE7EVAkqyLtiZfcCvG/h5GvTWX7+miZz4IgAErB7zVOkl/rpuBXHclAwi8BdDmAeBArh+KU4kFMJgEgaMArMSaBEBEenAX/++r6pN+eNS7dvC35/z4GWBv6tcv92PLjVehqo+q6kFVPTgyMlLPsWw8KRdQoNAd+o4a+aD5GEB+cYot2QwnRydRJVUIrgBdBpLHQmmP/kXOTcylCsH5tY0ZC6A/Oaadg/lKABhcGmi7p4AGss7S2eljHYXEAjCyMNsg1pIFJMB3gLdU9Rupp54GQibPYeCp1PjnfTbQdcC4dxU9C9wsIsM++HuzH2sfEhdQxd0zENIGrZiaIf/dp4L+9H3XHctMKeiAv8Dv3VJmdKKYKgVtpBtYwAeBAf7opl/k9z99VeW5WWsWwGSNBTBtZ2G2QaylrdOngM8Br4nIK37sa8DDwOMicjfwDnCnf+4Z4HbgFDADfAFAVcdE5OvAi/51D6rqWEOOYrPwLqC0BdAvIQZg5IMWLvBzE4z053j1tPP7mykFHfBCtztfZra8wJmLPne8OAmSsRPT8W0hUeXQ/lTIbXHBufTMWAB9UJpih48BDOWAxfloAazCqgKgqj8Clium8dklXq/APcv8raPA0Xom2FJ4F9D2/hyZLmFhUenrCn1HjQhAth+kK9kNXFpYBGBrfxAAAwFgSFxAO3MloIufnZviwEeG7XQDC/hSycwXoSfV4nJuHFAbWUDgjrM0xdW7BujLZrh6q3duWPlebhAGnLmbSH4QyjNkdJ4dPpjmKg5ixwIQSeoB7UgFDM20gwz44xjpqRSDK+S7fR0gA7uAA+FYSlPV41bqAAV8EHjPUC9vPHgrv7TNr22tWHIbRBSAeki5R0JaXS9zzmWQafPNNGlSFUEDw1usWQDOx7+1u9IQJikFbaEOUCApCDdRPW6lEmjAB4ETSm6Xd3QBrUwUgHoIqylfDgJ8P+Bsnx2XASTB7iAAhXw3PZkuWzEA/14Ohp2/YKsZTCBXKQldRSgEZ8oCSB1j2afzRhfQikQBqIckQ2aCy4Z6yWa6XDtIa2ZmrlIRFGBbfw4W5t0Ky4oA+PcytzDNQM65CwqWmsEEEgtgGReQhY1gkASBk4p3iQUQBWAlogDUQ+ICGufuG67k25874JrBWPuQhYqg/c7K2dqXrbgQLOwCBt8oJAPFCXb41MEBS81gAtlKT4AqrFQCDWT7XdbPgs/KSyyA6AJaiSgA9ZC4gJwFcNPHdkLZUEP4QH6wKgawtS9bcRlYsQBE3IXel4MAKhvBLAlAbhkBmBlzAmjl/ay1dKIFsCaiANRDygWUUJo26AIqQHGcbf1ZRFJ7AMDOBQPc+1msCEDFAjAYBC4tYQH0DtmJXYX2l+E4yz62Y21x1mCiANRDygWUYNIF5CyAHoE79u/hM1eP2BSA3CAUJ5N018GswvycMQFYJghsqQ4QpI7TWwDBBWTtu9lg1rITOBJIuYASyjOwZVtz5rNR5AuAQmmKb/72fjf2xgv+OUMC4NNdd+z1LqCkG5ghF1BPHyBLxwCspIBCqiR0rQvImHXeYKIFUA9dGWdqVrmAZux9yJaydOYu+ueMBIHBl0kY5/qrtnHgiiH2bpmvjFuhq6uqHlDC7AVbFkAIdgcBKHsBiC6gFYkCUC/eP55g0QW0lKVj0gXkLIBrLivw5B9+il7fItKUAEDSFKaKGUOVQOFSF1Bp2vUIyPQ0b05tQBSAevGbpBIsNp5e0gIYd1kjlo41X6i+MM4ZKwUd8H2Bq5g10g0skASBUxZA3AS2KlEA6sX7jRNMuoCWyHYKu4CtZI1AUtwv2TwUxMBSKQhIuoIllOd87MqSBbBEGmgsA7EqUQDqJVw0ABbKrtKitQ9a6IZV6wKy5P4Bd9FYnK+kDCa9AKwJQI0LKNkEZskCqPQFBmIvgDUSBaBe0i6gstFMg6VcQLMX7ewCDuRrYh3WuoEFQk+AgLVKoFDVFxhwom4tNrcBRAGol7QLKKSaWVtppLqCJZi0AIKlM1l9a04ACtVZQGFXtyUXECQloQEnBJbiVRtEFIB6SfuNy0ZLznbnoDtvXwBqYx3FSdcMx9rKsTYLyFodoEC6JHTZYGxuA4gCUC/5QVdwan7OrgsIqmMdYFMAksChF7pQB8hSoBt8EHiyEuxOKoEaE4B0rKNkMD17A4gCUC/pVaNVFxBcmu5qUgCWsAByxo4RnADoYmXBYtYC6EtZANEFtBZWFQAROSoi50Tk9dTYVhE5LiIn/e2wHxcR+ZaInBKRV0XkQOp3DvvXnxSRwxtzOJtAOkMmqTdi8IOWjnXMF2F+1tYuYEgFgVMxAGv+f7i0HtDMmFsdp3sEWyDbXwkCRwtgTazFAvgucGvN2H3Ac6q6D3jOPwa4Ddjnf44Aj4ATDOB+4JPAtcD9QTTajrQFENIHrbuALO4Chkt3PFsrBR3I1Qjd7AVbKaCBdBDY4gbNDWBVAVDVHwJjNcOHgGP+/jHgjtT499TxY2BIRHYDtwDHVXVMVS8Ax7lUVNqDJEXyYsoFZPCDlnYBJQJgzAIIF/sqF5BFAajpCWCtEmggBIEXF22WaNkA1hsD2KmqZ/3994Gd/v4e4N3U6077seXGL0FEjojICRE5cf78+XVObwNJrxotl5xNu4CsWgChuF/RuABka1xAsxfstIJME4LA83PuscXYXIP50EFgVVVAGzCX8PceVdWDqnpwZGSkUX+2cXSKC6jKAvCVQK1tBIMaV5dVF1CNBTBr1QLwQWCr6dkbwHoFYNS7dvC35/z4GWBv6nWX+7HlxtuP9C7ZkuG+o7lBF/idL7ldwGDPAoBqS6c4aa8OEKS6gnn/+IyxXgCB0Bc4bHSzuDBrMOsVgKeBkMlzGHgqNf55nw10HTDuXUXPAjeLyLAP/t7sx9qPbL/bLFSccCsNyUAm2+xZNZ58KtvJqgsIKpUyF8pO8KzVAYJqC0DVXi+AQDjOKb8ejS6gVVm1I5iI/AC4EdguIqdx2TwPA4+LyN3AO8Cd/uXPALcDp4AZ4AsAqjomIl8HXvSve1BVawPL7UGqmThdGef/t7ZxCKrLQZgWgIK7IFotAwEpAfBirgt2LQCAqVF3G11Aq7KqAKjqXcs89dklXqvAPcv8naPA0bpm16oE/3h3zu4qIx3snht3Vk63sbxxcEJ38X9tC0B3Hrq6XYqkxUqggbDfIVoAayb2BF4PuUFfD6jfZgYQVMc65i66FFCLlk4IAlsWAJFKPaDgH7foAgqxuGgBrJkoAOshBA4tFg4LpLOdLJaBCAR3ntVS0IGc7342Y7QSKFT6Ak9HC2CtxFpA6yH0BS4ZbjpRZQEYFoC8z3YKBdIs1gKCSlcwq3WAIOUC8vuHrC7OGkgUgPUQYgCWm07UxgCsCkA4zon3/GOrFoDf8Ga1EihcGgS2mJ7dYKIArIfgAipPGxcA6QALIAjAaXdrVgAGUkFgsfl+1qaBxn0AqxIFYD2EwKFlF1BXV8U/brEdZCBcNMbPVD+2RmgLOTPm3suuTLNn1HjCij/EALqjAKxGFID1kB909dWnP7BrAUDF1WXZAkhcQGdcUN+q2yCdBWQxBRQqfYEX5933site3lYjnqH1kG4mblkAcgWYeh8Wy3YFILyX42dsdgML5AqVILDFAHAgBIItfy8bSBSA9ZAuF2DVBQR+k5Qv4mpVAMJ7OfmezTIQgZAFNP2BzQBwIKSCWv5eNpAoAOshfTG0vNkkPwjjHSIAi/N2/f9QWRmPv9shFoDh72UDiQKwHqoEwHCgKVeo1Fa31gwmkK7+aVoA/LHNXjBuAfgLf7QA1kQUgPXQMS6glNBZFYDuHGRy7n4nCADYtgCyMQZQD1EA1kN61Wj5g5Y+TqsuIKgcp2UByKYFwKiYQwwC10kUgPVQ5QIy/EGrsgAMC0C48FsPAgdMu4BiELgeogCsh+68yzcG2x+0XIdYALkOsADCyhhsu4BiELguogCsB5GK28C0BZA6xm6DXc8CiQsoWgBtTwwC10UUgPUSVsSmBWCw+tYqHWEBpMTNsgUQg8B1EQVgvYQvlNXSAVApjRwFoP3JplxAli2A8B5a/l42kE0XABG5VUT+S0ROich9m/3/G0biHjG8D6BTLIBOyALqzrp010zO9uo4WgB1sakCICIZ4G+A24BrgLtE5JrNnEPD6AgXkL8wWheAYAHkDccAwAlc77DdekeQCgIbXpg1kM22AK4FTqnq/6hqCXgMOLTJc2gMwT1i2dRMLADDeePQGUFgcBdHy+4fSAWBDX8vG8hmC8Ae4N3U49N+rP3IF1z54Izh7JiQ7toJK2Oo9pNbJDdgOwAMlX0Ali3zBtJyTeFF5AhwBOCKK65o8mxW4FfugsG9ts1pEbjlIdh7bbNnsrFcfRvc8DZs39fsmWwsn/6K/SYpl+2HX/tjuOozzZ5JWyCqunn/TOR64AFVvcU//iqAqv75Uq8/ePCgnjhxYtPmF4lEIhYQkZdU9eBqr9tsF9CLwD4RuVJEssDvAE9v8hwikUgkwia7gFR1XkS+BDwLZICjqvrGZs4hEolEIo5NjwGo6jPAM5v9fyORSCRSTdwJHIlEIh1KFIBIJBLpUKIARCKRSIcSBSASiUQ6lCgAkUgk0qFs6kawehGR88A7H+JPbAc+aNB0Noo4x8YQ59gY4hwbQ7Pn+BFVHVntRS0tAB8WETmxlt1wzSTOsTHEOTaGOMfG0A5zhOgCikQikY4lCkAkEol0KNYF4NFmT2ANxDk2hjjHxhDn2BjaYY62YwCRSCQSWR7rFkAkEolElsGkALRD43kReVtEXhORV0SkZZoeiMhRETknIq+nxraKyHEROelvh1twjg+IyBl/Pl8RkdubPMe9IvK8iLwpIm+IyL1+vGXO5QpzbJlzKSJ5EfkPEfmJn+Of+fErReQF/x3/B19evtXm+F0R+XnqPO5v1hyXw5wLyDee/2/gN3AtJ18E7lLVN5s6sRpE5G3goKq2VD6ziPw6MAV8T1U/7sf+AhhT1Ye9oA6r6p+22BwfAKZU9S+bNa80IrIb2K2qL4vIAPAScAfwe7TIuVxhjnfSIudSRAToU9UpEekBfgTcC/wJ8KSqPiYi3wZ+oqqPtNgcvwj8i6r+YzPmtRYsWgB2Gs83AVX9ITBWM3wIOObvH8NdJJrGMnNsKVT1rKq+7O9PAm/h+l+3zLlcYY4tgzqm/MMe/6PATUC4sDb7PC43x5bHogC0S+N5Bf5VRF7yfZBbmZ2qetbffx/Y2czJrMCXRORV7yJqqpsqjYh8FPgE8AItei5r5ggtdC5FJCMirwDngOPAz4CLqjrvX9L073jtHFU1nMeH/Hn8pojkmjjFJbEoAO3CDap6ALgNuMe7NVoedT7DVlzdPAL8ArAfOAv8VXOn4xCRfuAJ4MuqOpF+rlXO5RJzbKlzqaoLqrofuBxn4X+smfNZito5isjHga/i5vqrwFagaW7T5bAoAGeAvanHl/uxlkJVz/jbc8A/4T7Yrcqo9xcHv/G5Js/nElR11H8JF4G/pQXOp/cHPwF8X1Wf9MMtdS6XmmMrnksAVb0IPA9cDwyJSOho2DLf8dQcb/UuNlXVIvB3tMh5TGNRAFq+8byI9PmgGyLSB9wMvL7ybzWVp4HD/v5h4KkmzmVJwkXV85s0+Xz6wOB3gLdU9Rupp1rmXC43x1Y6lyIyIiJD/n4vLrnjLdxF9rf8y5p9Hpea409TQi+4GEXLfcfNZQEB+LS1v6bSeP6hJk+pChG5CrfqB9eX+e9bZY4i8gPgRlw1w1HgfuCfgceBK3DVWe9U1aYFYZeZ4404l4UCbwN/kPK1bzoicgPw78BrwKIf/hrOx94S53KFOd5Fi5xLEfllXJA3g1uwPq6qD/rv0GM418p/Ar/rV9qtNMd/A0YAAV4BvpgKFrcEJgUgEolEIqtj0QUUiUQikTUQBSASiUQ6lCgAkUgk0qFEAYhEIpEOJQpAJBKJdChRACKRSKRDiQIQiUQiHUoUgEgkEulQ/h9+9eCWzIb2dQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11e74e518>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "pred2=pd.DataFrame(prediction2)\n",
    "pred3=pd.concat([pred2,y],axis=1)\n",
    "#pred3.head()\n",
    "\n",
    "plt.plot(pred3.iloc[:50,0]   ) \n",
    "plt.plot(pred3.iloc[:50,1]   )\n",
    "\n",
    "#plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 427,
   "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>0</th>\n",
       "      <th>Sales</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>62498.851562</td>\n",
       "      <td>60436</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>57663.335938</td>\n",
       "      <td>52848</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>57345.535156</td>\n",
       "      <td>55606</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>41966.773438</td>\n",
       "      <td>43727</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1743.363525</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>70284.820312</td>\n",
       "      <td>89391</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>69647.234375</td>\n",
       "      <td>72552</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>69359.898438</td>\n",
       "      <td>64297</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>69044.242188</td>\n",
       "      <td>64112</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>68702.062500</td>\n",
       "      <td>65006</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>41728.242188</td>\n",
       "      <td>43300</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>1537.195557</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               0  Sales\n",
       "0   62498.851562  60436\n",
       "1   57663.335938  52848\n",
       "2   57345.535156  55606\n",
       "3   41966.773438  43727\n",
       "4    1743.363525      0\n",
       "5   70284.820312  89391\n",
       "6   69647.234375  72552\n",
       "7   69359.898438  64297\n",
       "8   69044.242188  64112\n",
       "9   68702.062500  65006\n",
       "10  41728.242188  43300\n",
       "11   1537.195557      0"
      ]
     },
     "execution_count": 427,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred3.iloc[:12,]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}