Data Modeling - VAR.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "a6d894da",
   "metadata": {},
   "source": [
    "# Drilling Activity Prediction: Data Modeling - VAR\n",
    "\n",
    "- Load datasets and resample to monthly frequency\n",
    "- Use oil dataset to make predictions in oil rig count variable:\n",
    "    - Train/test split\n",
    "    - VAR Model Hyperparameter selection\n",
    "    - Fit the model\n",
    "    - Invert Transformation\n",
    "    - Generate forecast values and plot results\n",
    "    - Evaluation Metrics\n",
    "    - Fit the model again but with 100% of the data\n",
    "    - Forecast real future values\n",
    "\n",
    "\n",
    "- Create Model with Gas dataset to make predictions in Gas Rig Counts"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "3ac2f5f3",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "\n",
    "# Load specific forecasting tools\n",
    "from statsmodels.tsa.api import VAR\n",
    "from statsmodels.tsa.stattools import adfuller, grangercausalitytests\n",
    "from statsmodels.tools.eval_measures import rmse, mse, meanabs\n",
    "from sklearn.model_selection import TimeSeriesSplit\n",
    "import time\n",
    "from pylab import rcParams\n",
    "import datetime\n",
    "\n",
    "# Ignore harmless warnings\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")\n",
    "# Size of all plots\n",
    "rcParams['figure.figsize'] = 6,6"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "e11ea7d9",
   "metadata": {},
   "outputs": [],
   "source": [
    "dfoil = pd.read_csv('dfoil.csv',index_col=0,parse_dates=True)\n",
    "dfoil_transformed = pd.read_csv('dfoil_transformed.csv',index_col=0,parse_dates=True)\n",
    "\n",
    "\n",
    "dfgas = pd.read_csv('dfgas.csv',index_col=0,parse_dates=True)\n",
    "dfgas_transformed = pd.read_csv('dfgas_transformed.csv',index_col=0,parse_dates=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b0a86a26",
   "metadata": {},
   "source": [
    "A $K$-dimensional VAR model of order $p$, denoted <strong>VAR(p)</strong>, considers each variable $y_K$ in the system.<br>\n",
    "\n",
    "For example, The system of equations for a 2-dimensional VAR(1) model is:\n",
    "\n",
    "&nbsp;&nbsp;&nbsp;&nbsp;$y_{1,t} = c_1 + \\phi_{11,1}y_{1,t-1} + \\phi_{12,1}y_{2,t-1} + \\varepsilon_{1,t}$<br>\n",
    "&nbsp;&nbsp;&nbsp;&nbsp;$y_{2,t} = c_2 + \\phi_{21,1}y_{1,t-1} + \\phi_{22,1}y_{2,t-1} + \\varepsilon_{2,t}$\n",
    "\n",
    "where the coefficient $\\phi_{ii,l}$ captures the influence of the $l$th lag of variable $y_i$ on itself,<br>\n",
    "the coefficient $\\phi_{ij,l}$ captures the influence of the $l$th lag of variable $y_j$ on $y_i$,<br>\n",
    "and $\\varepsilon_{1,t}$ and $\\varepsilon_{2,t}$ are white noise processes that may be correlated.<br>\n",
    "\n",
    "Carrying this further, the system of equations for a 2-dimensional VAR(3) model is:\n",
    "\n",
    "&nbsp;&nbsp;&nbsp;&nbsp;$y_{1,t} = c_1 + \\phi_{11,1}y_{1,t-1} + \\phi_{12,1}y_{2,t-1} + \\phi_{11,2}y_{1,t-2} + \\phi_{12,2}y_{2,t-2} + \\phi_{11,3}y_{1,t-3} + \\phi_{12,3}y_{2,t-3} + \\varepsilon_{1,t}$<br>\n",
    "&nbsp;&nbsp;&nbsp;&nbsp;$y_{2,t} = c_2 + \\phi_{21,1}y_{1,t-1} + \\phi_{22,1}y_{2,t-1} + \\phi_{21,2}y_{1,t-2} + \\phi_{22,2}y_{2,t-2} + \\phi_{21,3}y_{1,t-3} + \\phi_{22,3}y_{2,t-3} + \\varepsilon_{2,t}$<br><br>\n",
    "\n",
    "and the system of equations for a 3-dimensional VAR(2) model is:\n",
    "\n",
    "&nbsp;&nbsp;&nbsp;&nbsp;$y_{1,t} = c_1 + \\phi_{11,1}y_{1,t-1} + \\phi_{12,1}y_{2,t-1} + \\phi_{13,1}y_{3,t-1} + \\phi_{11,2}y_{1,t-2} + \\phi_{12,2}y_{2,t-2} + \\phi_{13,2}y_{3,t-2} + \\varepsilon_{1,t}$<br>\n",
    "&nbsp;&nbsp;&nbsp;&nbsp;$y_{2,t} = c_2 + \\phi_{21,1}y_{1,t-1} + \\phi_{22,1}y_{2,t-1} + \\phi_{23,1}y_{3,t-1} + \\phi_{21,2}y_{1,t-2} + \\phi_{22,2}y_{2,t-2} + \\phi_{23,2}y_{3,t-2} + \\varepsilon_{2,t}$<br>\n",
    "&nbsp;&nbsp;&nbsp;&nbsp;$y_{3,t} = c_3 + \\phi_{31,1}y_{1,t-1} + \\phi_{32,1}y_{2,t-1} + \\phi_{33,1}y_{3,t-1} + \\phi_{31,2}y_{1,t-2} + \\phi_{32,2}y_{2,t-2} + \\phi_{33,2}y_{3,t-2} + \\varepsilon_{3,t}$<br><br>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "b77b6fa6",
   "metadata": {},
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "This cell is used only to Test model in different points in time\n",
    "\"\"\"\n",
    "# dfoil = dfoil[:'2016']\n",
    "# dfoil_transformed = dfoil_transformed[:'2016']\n",
    "# dfgas = dfgas[:'2016']\n",
    "# dfgas_transformed = dfgas[:'2016']\n",
    "freq_time = 'M'\n",
    "dfoil = dfoil.resample('M').mean()\n",
    "dfoil.index.freq = 'M'\n",
    "dfoil_transformed = dfoil_transformed.resample('M').mean()\n",
    "dfoil_transformed.index.freq = 'M'\n",
    "dfgas = dfgas.resample('M').mean()\n",
    "dfgas.index.freq = 'M'\n",
    "dfgas_transformed = dfgas_transformed.resample('M').mean()\n",
    "dfgas_transformed.index.freq = 'M'"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bb99a92b",
   "metadata": {},
   "source": [
    "#### Train/test split\n",
    "\n",
    "- We will evaluate the model forecasting 12 months ahead (1 year)\n",
    "- Train test split. Test df will have 12 data points"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "fadf9715",
   "metadata": {},
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "nobs is the variable that will be used to determine the number of months to predict.\n",
    "\"\"\"\n",
    "\n",
    "nobs=12\n",
    "train, test = dfoil_transformed[0:-nobs], dfoil_transformed[-nobs:]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a30d0047",
   "metadata": {},
   "source": [
    "#### Fit VAR model and hyperparameter selection"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "0d47e041",
   "metadata": {},
   "outputs": [],
   "source": [
    "oil_model = VAR(train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "b4731be9",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Order = 0\n",
      "AIC:  25.85377895509131\n",
      "BIC:  26.04552677733711\n",
      "\n",
      "Order = 1\n",
      "AIC:  24.625627487008828\n",
      "BIC:  25.13825754739741\n",
      "\n",
      "Order = 2\n",
      "AIC:  24.543052490949258\n",
      "BIC:  25.378205045457985\n",
      "\n",
      "Order = 3\n",
      "AIC:  24.49046824264916\n",
      "BIC:  25.649797644726515\n",
      "\n",
      "Order = 4\n",
      "AIC:  24.44370500807985\n",
      "BIC:  25.928879877434046\n",
      "\n",
      "Order = 5\n",
      "AIC:  24.47985186250436\n",
      "BIC:  26.29255525647847\n",
      "\n",
      "Order = 6\n",
      "AIC:  24.51040237882233\n",
      "BIC:  26.652331966404212\n",
      "\n",
      "Order = 7\n",
      "AIC:  24.560680628038043\n",
      "BIC:  27.033548866554383\n",
      "\n",
      "Order = 8\n",
      "AIC:  24.662155460325025\n",
      "BIC:  27.467689774869942\n",
      "\n",
      "Order = 9\n",
      "AIC:  24.578373424853176\n",
      "BIC:  27.718316390502856\n",
      "\n"
     ]
    }
   ],
   "source": [
    "for i in range(10):\n",
    "    oil_results = oil_model.fit(i,trend='ctt')\n",
    "    print('Order =', i)\n",
    "    print('AIC: ', oil_results.aic)\n",
    "    print('BIC: ', oil_results.bic)\n",
    "    print()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "b3e1d6e4",
   "metadata": {},
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "Choose best AIC based on test done before (4)\n",
    "\"\"\"\n",
    "\n",
    "oil_results = oil_model.fit(4,trend='ctt')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "473a7e8b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['oilrig_count', 'oil_price', 'oilconsumption', 'oilproduction', 'oilstorage']"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "oil_model.endog_names"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "e9ad768e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "  Summary of Regression Results   \n",
       "==================================\n",
       "Model:                         VAR\n",
       "Method:                        OLS\n",
       "Date:           Sun, 21, Nov, 2021\n",
       "Time:                     11:00:43\n",
       "--------------------------------------------------------------------\n",
       "No. of Equations:         5.00000    BIC:                    25.9289\n",
       "Nobs:                     282.000    HQIC:                   25.0393\n",
       "Log likelihood:          -5332.27    FPE:                4.13575e+10\n",
       "AIC:                      24.4437    Det(Omega_mle):     2.79448e+10\n",
       "--------------------------------------------------------------------\n",
       "Results for equation oilrig_count\n",
       "====================================================================================\n",
       "                       coefficient       std. error           t-stat            prob\n",
       "------------------------------------------------------------------------------------\n",
       "const                    -0.387508         0.281652           -1.376           0.169\n",
       "trend                     0.007407         0.004572            1.620           0.105\n",
       "trend**2                 -0.000025         0.000015           -1.613           0.107\n",
       "L1.oilrig_count           0.858716         0.066833           12.849           0.000\n",
       "L1.oil_price              0.950114         0.251206            3.782           0.000\n",
       "L1.oilconsumption         0.001623         0.002392            0.679           0.497\n",
       "L1.oilproduction         -0.009210         0.007035           -1.309           0.190\n",
       "L1.oilstorage            -0.000219         0.000122           -1.793           0.073\n",
       "L2.oilrig_count          -0.329181         0.086739           -3.795           0.000\n",
       "L2.oil_price              0.755291         0.259553            2.910           0.004\n",
       "L2.oilconsumption        -0.003538         0.002560           -1.382           0.167\n",
       "L2.oilproduction         -0.008588         0.006919           -1.241           0.215\n",
       "L2.oilstorage             0.000151         0.000124            1.219           0.223\n",
       "L3.oilrig_count           0.148563         0.086438            1.719           0.086\n",
       "L3.oil_price              0.447634         0.262889            1.703           0.089\n",
       "L3.oilconsumption         0.000688         0.002556            0.269           0.788\n",
       "L3.oilproduction          0.002938         0.006827            0.430           0.667\n",
       "L3.oilstorage             0.000046         0.000125            0.366           0.714\n",
       "L4.oilrig_count           0.000475         0.062722            0.008           0.994\n",
       "L4.oil_price             -0.011158         0.253895           -0.044           0.965\n",
       "L4.oilconsumption         0.002900         0.002318            1.251           0.211\n",
       "L4.oilproduction         -0.002812         0.006861           -0.410           0.682\n",
       "L4.oilstorage            -0.000063         0.000125           -0.501           0.617\n",
       "====================================================================================\n",
       "\n",
       "Results for equation oil_price\n",
       "====================================================================================\n",
       "                       coefficient       std. error           t-stat            prob\n",
       "------------------------------------------------------------------------------------\n",
       "const                    -0.011037         0.073698           -0.150           0.881\n",
       "trend                     0.000721         0.001196            0.602           0.547\n",
       "trend**2                 -0.000003         0.000004           -0.867           0.386\n",
       "L1.oilrig_count          -0.008947         0.017488           -0.512           0.609\n",
       "L1.oil_price              0.245457         0.065732            3.734           0.000\n",
       "L1.oilconsumption         0.000270         0.000626            0.431           0.666\n",
       "L1.oilproduction         -0.000031         0.001841           -0.017           0.987\n",
       "L1.oilstorage            -0.000008         0.000032           -0.264           0.792\n",
       "L2.oilrig_count          -0.027058         0.022697           -1.192           0.233\n",
       "L2.oil_price              0.088878         0.067916            1.309           0.191\n",
       "L2.oilconsumption        -0.000270         0.000670           -0.403           0.687\n",
       "L2.oilproduction          0.002261         0.001810            1.249           0.212\n",
       "L2.oilstorage             0.000034         0.000032            1.057           0.291\n",
       "L3.oilrig_count           0.015210         0.022618            0.672           0.501\n",
       "L3.oil_price             -0.013675         0.068789           -0.199           0.842\n",
       "L3.oilconsumption        -0.000321         0.000669           -0.479           0.632\n",
       "L3.oilproduction          0.002659         0.001786            1.489           0.137\n",
       "L3.oilstorage            -0.000016         0.000033           -0.475           0.634\n",
       "L4.oilrig_count           0.001755         0.016412            0.107           0.915\n",
       "L4.oil_price              0.005917         0.066435            0.089           0.929\n",
       "L4.oilconsumption        -0.000325         0.000607           -0.536           0.592\n",
       "L4.oilproduction         -0.001987         0.001795           -1.107           0.268\n",
       "L4.oilstorage             0.000016         0.000033            0.502           0.615\n",
       "====================================================================================\n",
       "\n",
       "Results for equation oilconsumption\n",
       "====================================================================================\n",
       "                       coefficient       std. error           t-stat            prob\n",
       "------------------------------------------------------------------------------------\n",
       "const                     4.285288         7.455949            0.575           0.565\n",
       "trend                    -0.058551         0.121038           -0.484           0.629\n",
       "trend**2                  0.000127         0.000407            0.312           0.755\n",
       "L1.oilrig_count           1.682829         1.769211            0.951           0.342\n",
       "L1.oil_price              2.841539         6.649962            0.427           0.669\n",
       "L1.oilconsumption        -0.462268         0.063327           -7.300           0.000\n",
       "L1.oilproduction          0.344959         0.186240            1.852           0.064\n",
       "L1.oilstorage            -0.011982         0.003240           -3.698           0.000\n",
       "L2.oilrig_count          -3.531708         2.296174           -1.538           0.124\n",
       "L2.oil_price              4.903460         6.870931            0.714           0.475\n",
       "L2.oilconsumption        -0.349321         0.067769           -5.155           0.000\n",
       "L2.oilproduction          0.080606         0.183157            0.440           0.660\n",
       "L2.oilstorage            -0.001071         0.003284           -0.326           0.744\n",
       "L3.oilrig_count           2.506909         2.288189            1.096           0.273\n",
       "L3.oil_price             -1.707170         6.959237           -0.245           0.806\n",
       "L3.oilconsumption        -0.193521         0.067654           -2.860           0.004\n",
       "L3.oilproduction         -0.059255         0.180715           -0.328           0.743\n",
       "L3.oilstorage             0.009260         0.003306            2.801           0.005\n",
       "L4.oilrig_count           0.542791         1.660398            0.327           0.744\n",
       "L4.oil_price              7.255037         6.721154            1.079           0.280\n",
       "L4.oilconsumption        -0.093737         0.061370           -1.527           0.127\n",
       "L4.oilproduction         -0.464243         0.181618           -2.556           0.011\n",
       "L4.oilstorage             0.012768         0.003321            3.845           0.000\n",
       "====================================================================================\n",
       "\n",
       "Results for equation oilproduction\n",
       "====================================================================================\n",
       "                       coefficient       std. error           t-stat            prob\n",
       "------------------------------------------------------------------------------------\n",
       "const                    -2.404820         2.511925           -0.957           0.338\n",
       "trend                     0.053292         0.040778            1.307           0.191\n",
       "trend**2                 -0.000128         0.000137           -0.934           0.350\n",
       "L1.oilrig_count           0.928539         0.596051            1.558           0.119\n",
       "L1.oil_price             -0.514054         2.240387           -0.229           0.819\n",
       "L1.oilconsumption         0.036696         0.021335            1.720           0.085\n",
       "L1.oilproduction         -0.174304         0.062745           -2.778           0.005\n",
       "L1.oilstorage            -0.000071         0.001092           -0.065           0.948\n",
       "L2.oilrig_count           0.326274         0.773586            0.422           0.673\n",
       "L2.oil_price             -1.117492         2.314832           -0.483           0.629\n",
       "L2.oilconsumption        -0.018370         0.022832           -0.805           0.421\n",
       "L2.oilproduction         -0.038803         0.061706           -0.629           0.529\n",
       "L2.oilstorage            -0.003543         0.001106           -3.202           0.001\n",
       "L3.oilrig_count          -1.172175         0.770896           -1.521           0.128\n",
       "L3.oil_price             -1.478395         2.344582           -0.631           0.528\n",
       "L3.oilconsumption        -0.020212         0.022793           -0.887           0.375\n",
       "L3.oilproduction          0.159799         0.060883            2.625           0.009\n",
       "L3.oilstorage            -0.002226         0.001114           -1.999           0.046\n",
       "L4.oilrig_count           0.637692         0.559392            1.140           0.254\n",
       "L4.oil_price             -6.262839         2.264371           -2.766           0.006\n",
       "L4.oilconsumption         0.005785         0.020676            0.280           0.780\n",
       "L4.oilproduction         -0.019325         0.061187           -0.316           0.752\n",
       "L4.oilstorage            -0.002505         0.001119           -2.239           0.025\n",
       "====================================================================================\n",
       "\n",
       "Results for equation oilstorage\n",
       "====================================================================================\n",
       "                       coefficient       std. error           t-stat            prob\n",
       "------------------------------------------------------------------------------------\n",
       "const                   -47.920936       146.643740           -0.327           0.744\n",
       "trend                     3.237055         2.380573            1.360           0.174\n",
       "trend**2                 -0.012663         0.008000           -1.583           0.113\n",
       "L1.oilrig_count         -22.596006        34.796885           -0.649           0.516\n",
       "L1.oil_price           -164.119071       130.791586           -1.255           0.210\n",
       "L1.oilconsumption        -3.843109         1.245511           -3.086           0.002\n",
       "L1.oilproduction          0.922703         3.662963            0.252           0.801\n",
       "L1.oilstorage             0.121851         0.063729            1.912           0.056\n",
       "L2.oilrig_count         -34.369706        45.161187           -0.761           0.447\n",
       "L2.oil_price           -134.705388       135.137603           -0.997           0.319\n",
       "L2.oilconsumption        -2.084083         1.332886           -1.564           0.118\n",
       "L2.oilproduction         -2.615032         3.602329           -0.726           0.468\n",
       "L2.oilstorage            -0.107311         0.064592           -1.661           0.097\n",
       "L3.oilrig_count         -28.214238        45.004147           -0.627           0.531\n",
       "L3.oil_price            120.410354       136.874417            0.880           0.379\n",
       "L3.oilconsumption         0.021273         1.330618            0.016           0.987\n",
       "L3.oilproduction          4.451663         3.554311            1.252           0.210\n",
       "L3.oilstorage            -0.017485         0.065015           -0.269           0.788\n",
       "L4.oilrig_count          30.705884        32.656732            0.940           0.347\n",
       "L4.oil_price           -285.253535       132.191781           -2.158           0.031\n",
       "L4.oilconsumption        -1.424366         1.207023           -1.180           0.238\n",
       "L4.oilproduction          3.077314         3.572061            0.861           0.389\n",
       "L4.oilstorage            -0.171073         0.065312           -2.619           0.009\n",
       "====================================================================================\n",
       "\n",
       "Correlation matrix of residuals\n",
       "                  oilrig_count  oil_price  oilconsumption  oilproduction  oilstorage\n",
       "oilrig_count          1.000000   0.276767        0.260161       0.080324   -0.212661\n",
       "oil_price             0.276767   1.000000        0.017443      -0.091995   -0.220097\n",
       "oilconsumption        0.260161   0.017443        1.000000       0.259312   -0.157363\n",
       "oilproduction         0.080324  -0.091995        0.259312       1.000000    0.040058\n",
       "oilstorage           -0.212661  -0.220097       -0.157363       0.040058    1.000000\n",
       "\n"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "oil_results.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "b419eaeb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "4"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lag_order = oil_results.k_ar\n",
    "lag_order"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1f2f065f",
   "metadata": {},
   "source": [
    "#### Generate Forecast values and plot results\n",
    "\n",
    "- Create forecasts of differenced dataframe\n",
    "- Transform forecast results into dataframe\n",
    "- Add datetime index to forecast dataframe\n",
    "- Invert dataframe transformation\n",
    "- Plot results and compare Forecasted values Vs Test set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "f2c377b6",
   "metadata": {},
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "y: Grab the last 4 values (lagged values) before the test set starts in format p x k (series).\n",
    "steps: amount of steps you want to forecast\n",
    "\"\"\"\n",
    "\n",
    "z = oil_results.forecast(y=train.values[-lag_order:], steps=nobs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "69fb0c56",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Timestamp('2020-11-30 00:00:00', freq='M')"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test.index.min()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "fca09eed",
   "metadata": {},
   "outputs": [],
   "source": [
    "idx = pd.date_range(test.index.min(), periods=nobs, freq='M')\n",
    "dfoil_forecast = pd.DataFrame(z, index=idx, columns=['oilrig_count2d', 'oil_price2d', 'oilconsumption2d', 'oilproduction2d',\n",
    "       'oilstorage2d'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "bc61d086",
   "metadata": {},
   "outputs": [],
   "source": [
    "dfoil_forecast['oilrig_count1d'] = (dfoil['oilrig_count'].iloc[-nobs-1]-dfoil['oilrig_count'].iloc[-nobs-2])+dfoil_forecast['oilrig_count2d'].cumsum()\n",
    "dfoil_forecast['oilrig_forecast'] = dfoil['oilrig_count'].iloc[-nobs-1] + dfoil_forecast['oilrig_count1d'].cumsum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "1ddd3c94",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:title={'center':'Oil Rig Count Prediction: FORECAST vs TEST Set - VAR Model'}>"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "dfoil['oilrig_count'].loc['2017':].plot(legend=True)\n",
    "dfoil_forecast['oilrig_forecast'].plot(legend=True,linewidth=4,title='Oil Rig Count Prediction: FORECAST vs TEST Set - VAR Model')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1a8711ef",
   "metadata": {},
   "source": [
    "#### Evaluation Metrics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "7c2bfea6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Oil Rig VAR(4) RMSE: 11.013\n",
      "Oil Rig VAR(4) MSE: 121.297\n",
      "Oil Rig VAR(4) MAE: 9.293\n",
      "\n",
      "Oil Rig Mean Value: 345.055\n",
      "Oil Rig std Value: 54.791\n",
      "Oil Rig Percent Change: 3.192%\n"
     ]
    }
   ],
   "source": [
    "RMSE1 = rmse(dfoil['oilrig_count'].loc[test.index.min():],dfoil_forecast['oilrig_forecast'])\n",
    "MSE1 = mse(dfoil['oilrig_count'].loc[test.index.min():],dfoil_forecast['oilrig_forecast'])\n",
    "MAE1 = meanabs(dfoil['oilrig_count'].loc[test.index.min():],dfoil_forecast['oilrig_forecast'])\n",
    "\n",
    "print(f\"\"\"Oil Rig VAR(4) RMSE: {RMSE1:.3f}\n",
    "Oil Rig VAR(4) MSE: {MSE1:.3f}\n",
    "Oil Rig VAR(4) MAE: {MAE1:.3f}\\n\n",
    "Oil Rig Mean Value: {dfoil['oilrig_count'][-nobs:].mean():.3f}\n",
    "Oil Rig std Value: {dfoil['oilrig_count'][-nobs:].std():.3f}\n",
    "Oil Rig Percent Change: {RMSE1/dfoil['oilrig_count'][-nobs:].mean()*100:.3f}%\"\"\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "0035e47f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "TRAIN: 178 TEST: 12\n",
      "TRAIN: 190 TEST: 12\n",
      "TRAIN: 202 TEST: 12\n",
      "TRAIN: 214 TEST: 12\n",
      "TRAIN: 226 TEST: 12\n",
      "TRAIN: 238 TEST: 12\n",
      "TRAIN: 250 TEST: 12\n",
      "TRAIN: 262 TEST: 12\n",
      "TRAIN: 274 TEST: 12\n",
      "TRAIN: 286 TEST: 12\n",
      "\n",
      "10-fold Cross Validation Oil Rig\n",
      "RMSE: 515.152,\n",
      "MSE: 382589.856\n",
      "MAE: 488.906\n",
      "\n",
      "Oil Rig Mean Value: 345.055\n",
      "Oil Rig std Value: 54.791\n",
      "Oil Rig Percent Change: 149.296%\n"
     ]
    }
   ],
   "source": [
    "tscv = TimeSeriesSplit(n_splits=10, test_size=12)\n",
    "RMSE_oil = []\n",
    "MSE_oil = []\n",
    "MAE_oil = []\n",
    "for train_index, test_index in tscv.split(dfoil_transformed):\n",
    "    print(\"TRAIN:\", len(train_index), \"TEST:\", len(test_index))\n",
    "    train,test = dfoil_transformed.loc[dfoil_transformed.index[train_index],:], dfoil_transformed.loc[dfoil_transformed.index[test_index],:]\n",
    "    oil_model = VAR(train)\n",
    "    oil_results = oil_model.fit(4,trend='ctt')\n",
    "    z = oil_results.forecast(y=train.values[-lag_order:], steps=nobs)\n",
    "    idx = pd.date_range(test.index.min(), periods=nobs, freq='M')\n",
    "    dfoil_forecast = pd.DataFrame(z, index=idx, columns=['oilrig_count2d', 'oil_price2d', 'oilconsumption2d', 'oilproduction2d',\n",
    "           'oilstorage2d'])\n",
    "    dfoil_forecast['oilrig_count1d'] = (dfoil['oilrig_count'].iloc[-nobs-1]-dfoil['oilrig_count'].iloc[-nobs-2])+dfoil_forecast['oilrig_count2d'].cumsum()\n",
    "    dfoil_forecast['oilrig_forecast'] = dfoil['oilrig_count'].iloc[-nobs-1] + dfoil_forecast['oilrig_count1d'].cumsum()\n",
    "    RMSE_oil.append('{:.3f}'.format(rmse(dfoil.loc[dfoil.index[test_index],'oilrig_count'],dfoil_forecast['oilrig_forecast'])))\n",
    "    MSE_oil.append('{:.3f}'.format(mse(dfoil.loc[dfoil.index[test_index],'oilrig_count'],dfoil_forecast['oilrig_forecast'])))\n",
    "    MAE_oil.append('{:.3f}'.format(meanabs(dfoil.loc[dfoil.index[test_index],'oilrig_count'],dfoil_forecast['oilrig_forecast'])))\n",
    "\n",
    "RMSE_oil = np.array(RMSE_oil).astype('float')\n",
    "MSE_oil = np.array(MSE_oil).astype('float')\n",
    "MAE_oil = np.array(MAE_oil).astype('float')\n",
    "print(f\"\"\"\\n10-fold Cross Validation Oil Rig\n",
    "RMSE: {RMSE_oil.mean():.3f},\n",
    "MSE: {MSE_oil.mean():.3f}\n",
    "MAE: {MAE_oil.mean():.3f}\\n\n",
    "Oil Rig Mean Value: {dfoil['oilrig_count'][-nobs:].mean():.3f}\n",
    "Oil Rig std Value: {dfoil['oilrig_count'][-nobs:].std():.3f}\n",
    "Oil Rig Percent Change: {RMSE_oil.mean()/dfoil['oilrig_count'][-nobs:].mean()*100:.3f}%\"\"\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "743a8a58",
   "metadata": {},
   "source": [
    "#### Fit Model with 100% of Data and forecast Real Future"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "669a5c17",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Execution time in seconds: 0.015956401824951172\n"
     ]
    }
   ],
   "source": [
    "startTime = time.time()\n",
    "oil_model = VAR(dfoil_transformed)\n",
    "oil_results = oil_model.fit(4,trend='ctt')\n",
    "z = oil_results.forecast(y=dfoil_transformed.values, steps=nobs)\n",
    "idx = pd.date_range(dfoil_transformed.index.max(), periods=nobs, freq='M')\n",
    "dfoil_forecast = pd.DataFrame(z, index=idx, columns=['oilrig_count2d', 'oil_price2d', 'oilconsumption2d', 'oilproduction2d',\n",
    "       'oilstorage2d'])\n",
    "dfoil_forecast['oilrig_count1d'] = (dfoil['oilrig_count'].iloc[-1]-dfoil['oilrig_count'].iloc[-2])+dfoil_forecast['oilrig_count2d'].cumsum()\n",
    "dfoil_forecast['oilrig_forecast'] = dfoil['oilrig_count'].iloc[-1] + dfoil_forecast['oilrig_count1d'].cumsum()\n",
    "executionTime = (time.time() - startTime)\n",
    "print('Execution time in seconds: ' + str(executionTime))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "e106fc1b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:title={'center':'Oil Rig Count Prediction: REAL FORECAST - VAR Model'}>"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "dfoil['oilrig_count'].loc['2017':].plot(legend=True)\n",
    "dfoil_forecast['oilrig_forecast'].plot(legend=True,linewidth=3,title='Oil Rig Count Prediction: REAL FORECAST - VAR Model')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a402e62b",
   "metadata": {},
   "source": [
    "#### Create model with Gas Rig Counts"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "30d3853c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['gasrig_count', 'gas_price', 'gaswell_count', 'gasimports_amount',\n",
       "       'gasconsumption_amount'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dfgas.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "d0d1b918",
   "metadata": {},
   "outputs": [],
   "source": [
    "train, test = dfgas_transformed[0:-nobs], dfgas_transformed[-nobs:]\n",
    "gas_model = VAR(train)\n",
    "gas_results = gas_model.fit(4,trend='ctt')\n",
    "z2 = gas_results.forecast(y=train.values, steps=nobs)\n",
    "idx2 = pd.date_range(test.index.min(), periods=nobs, freq='M')\n",
    "dfgas_forecast = pd.DataFrame(z2, index=idx2, columns=['gasrig_count2d', 'gas_price2d', 'gaswell_count2d', 'gasimports_amount2d',\n",
    "       'gasconsumption_amount2d'])\n",
    "dfgas_forecast['gasrig_count1d'] = (dfgas['gasrig_count'].iloc[-nobs-1]-dfgas['gasrig_count'].iloc[-nobs-2])+dfgas_forecast['gasrig_count2d'].cumsum()\n",
    "dfgas_forecast['gasrig_forecast'] = dfgas['gasrig_count'].iloc[-nobs-1] + dfgas_forecast['gasrig_count1d'].cumsum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "8db201e2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYkAAAF1CAYAAAAdjRbfAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/MnkTPAAAACXBIWXMAAAsTAAALEwEAmpwYAABRKUlEQVR4nO3dd3xV9fnA8c+TQUJ2QgYJCQkbwiZhD0FQwIWiddQBdVftsK111LZaqz/baqdaxWoRN4rbumWIAhLZewYSQiAQsggJGd/fH+cELyE7N7kjz/v1uq/cnHPuOc85997z3O845yvGGJRSSqm6+Lg6AKWUUu5Lk4RSSql6aZJQSilVL00SSiml6qVJQimlVL00SSillKqXVyUJEXlaRH7r6jg8jYgYEeltP2/xMRSREhHp6dzolFJNJSKTRSS7ics+ICIvNbacU5KEiFwpIqtE5LiIHLaf3yYi4oz1O2wnU0RO2CejXBGZLyIhNfONMbcaYx5qxfqni8gyESkWkTwRWSoiFzkn+ga3myki0xqYP1lEqu39LhaR7SLyo7aIpanHUESWiMiNtV4bYozZ0xZx1dq24+eg5pFgzwsQkf8Tkf32MjtF5C7Hz6Ide5n9uiMi8paIxDvMf0BEKmqtv8BhvojIT0Vkk/2ZzxaRN0RkcK04H7AT8Oha0zuJyOP260rs/fm7Pc9xm9W19vPqNjiWVzus/4TD56xEREoaON5POHtfRCRRRBbZ70mhfXznNnE/zvg8NvM4fCwif6hj+iz7XONn/z/Zfk/vrrVcij29xOE43NPINo1Y50s/h2n+9jS3uYCt1UlCRH4J/AP4C9AViANuBcYDnVq7/jpcaIwJAYYBw4F7nbFSEbkMeANYACRi7cfvgAudsX4nyLH3Owy4G3hWRFJrL+T4gfNyF9pJqeaRY09/A5gKnAeEAtcCN2N9Rh3dYR/P3kAI8Fit+a/XWn+Ew7x/AD8DfgpEAX2Bd4Dzaxawk9J1QL7919G9QDowyo5xMrAGTiXaEDu2/bX28+UmH50mMsa87LC9mdifM4dpNWof7zvaYF9eBLKAZKAL1nt3yNn7XI8XgGscf0zYrgVeNsZU2v/Poe73tEaEvb+XAb8VkXMa2e4xrONeY6Y9zX0YY1r8AMKB48CljSx3PrAWKML6EDzgMC8QeAk4ChQAq4G4etaTCUxz+P/PwIcO/88H/ujw/6+Bg0AOcCNggN51rFewPsR3NbAPPsD9wD7gMFYyCbfnTQay64sVeABYaL+mGNgMpNvzXgSqgRNACfDrOrZd1/rzsD6Ic4Gvgb/Zx/CPQADWSW8/1pfsaaCzw2vvcjgu1zselzqO4Sxgnf3e7QZmAA8DVUCZHfMT9rKO6wm39zfPPmb3Az72vLnAcjvGY8BeYGYzPnenfQ4cpk+1Y0qqNX20HW9NbEuAGx3m3wZsdvj/AeClerbdx17XqEZinGS/p1fb70snh3kfAD9v6X7W2q9cwNdh2iXABvv5KCDDfu8OAX9tZHtnfM4ai8NZ+2IvUwIMa2D+GOAbrPPEemCyPb3Oz2NzHkBnoBCY5DAt0l7nUPv/YKzv75XASezvsD0vxf78+zlM+5aGzynG/l684TDtTeA3gHGYlgC8h5WcdgE31Yp7vv092oL13c6u9dpFWN/DvcBPm/I5d3y0tiQxFuuE9G4jyx3HyrwRWAnjxyJysT1vDtYJJQnr18OtWF+uBolIIlbW3VXP/BnAL4BpWL8WJzewun729t9sYJm59mMK0BPr1+cTjcXp4CLgNaxj8F7Na40x13L6r6w/N7QSEfERkUvs9Wy0J48G9mCVfh4GHsX6dTsMa9+7YZWKao7Lr4BzsE54DVVzjcI60d9lb28SkGmM+Q3wFfavcfP9r0pH/8J6X3sCZ2G9/45VZKOB7UA0VrJ/ruZXnIjcIyIfNHQc6nEOsMoYk+U40RizCsjGSiK197ELMJt6Pkd1mIr1Jfy2keXmAO9j/TiA00ukK4FfiFUlO7iOX69NYu/XceBsh8k/BF6xn/8D+IcxJgzo5RCLMzllXxzW9aRY1dfdHWeISDfgQ6wfQVFYn+FFIhLTxM9jg4wxJ7COj2MJ4XJgmzFmvf3/bKwk9AbwCdZ7XCcRGQMMovHP1TvAJBGJEJFIYCJnnk9fw/r8JmD9MHxERGre899jvbe9gOmOMYmID9ZncD3WOWAq8HMRmd5ITKdrbsatlQmvAXJrTavJ9CdwyMq1lvk78Df7+fX2a4Y08ddICVY2N8AXWMW7mvnzsX8FA88D/+cwrzf1lyTG2/MCG9j2F8BtDv/3AyoAP5pWkvjcYV4qcKKuZevZ9mSs0kYB1q+JdcCV9ry5wH6HZQXrxNHLYdpYYK/DcXnUYV5f6ilJAM/UvE91xLQEh1/jDr+MegO+WL+0Uh3m3QIscYh5l8O8IPu1XZv4uav5HBTYj3fs6f8BXqvnNSuB3zjEXor1y9HYx7O7w7IP2PEXODwW2/N+A6xsJL4grF/vFzscx3cd5vsCt2OVAMuxSnRz6tnPxn59/xF43n4ear/3yfb/y4AHgegmHtfJ1F+ScDzeBdi/Zp28L5FYP3A2Y5UM1gEj7Xl3Ay/WWv6Tmm3V9Xls7gOYYO9boP3/18CdDvM/B/5uP78K69e5v/1/iv1ZKsA69xmskrI0sL2a78t/sL4ftwLP2tOMvUySfSxCHV73f8B8+/keYIbDvJtr3kOsH2L7a23zXuC/Dp/zNi9JHAWiHevBjTHjjFV/exS7zUNERovIYrEagwvtgxFtv+RFrDf7NRHJEZE/i4h/A9u82BhTU/fZ32E9tSVgVW3VyKpnuZr9AIhvYJkErGqTGvuwEkRcA69xlOvwvBQIbGb7QY4xJsIYE2WMGWaMec1hnuO+xWCdpL4TkQKxGlw/tqfX7Ifj8o77VFsSVhVTc0UD/px5vLo5/H/qeBhjSu2njnXgjbnYPh4RxpiL7WlHqP89jLfn1/ipMSYcGIJ1ckqstfxCh/VHGGOm2NOPNrCNGpcAlcD/7P9fBmaKSAyAMabKGPOkMWY8VgntYeB5ERnQyHrr8gowW0QCsH7prjHG1Bz3G7B+BGwTkdUickEL1l/j4lrH41ln74sx5pgx5h5jzECs79U64B27dJIM/KDmM21/rifQ+HsBnOq1V9OofF8921+O9Rm5WER6YVXXvWK/PgmrFqGmLeVdrKry82utJhrrc/xLrHNUQ+eyGguwSjDX2c8dJQD5xphih2mO36WGvs/JQEKtY3YfTT9nAa1vuF6B9ethViPLvYJVxZJkfzGfxvrFizGmwhjzoDEmFRgHXED9jUKnGGOWYv3qrd3gWOMgp3/xkxpY3XasA31pA8vkYB30Gt2xTgSHsH69BdXMEBFfvj8pN4VpxrKNvf4I1i+ZgQ5f6HDzfSPkQU4/FqcV62vJwirGNrbN2o5glbJqH68DDbzGGT4HRttf6FPE6l2UBHxZ+wXGmI1Yv8afbGJVyRdAooikN7DMHKwTxX4RycWqnvDHqgqqvf0TxpgnseqUz+iI0BhjzBasE8NMTq9qwhiz0xhzFRAL/Al4U0SCm7uNZsTSqn2pta4jWN/tBKzqpSyskoRjogo2xjxa85JG1ner+b7R/JEGFq05YV8DfGKMqWk4vxbrfPm+/Z7uwUoSZ1Q52Ynzr1jtGbc1YXe/wkp2cVhtdY5ygCgRCXWY5vhdauj7nIVVg+B4zEKNMec1IaZTWpUkjDEFWMXZp0TkMhEJtevMh2E18tQIxcqGZXY996kvi4hMsesyfbGK6BVYVStN8XfgHBEZWse8hcCPRGSAiAQB9fb9t8t1v8DqjfAjEQmz92OCiMyzF3sVuFNEeojV7fYRrB4wlcAOrJLB+XYp6H6stpqmOoRVd99qxphqrCLr30QkFqz6XId6yIXAXBFJtY/L7xtY3XNYx3CqfTy6iUj/xmI2xlTZ23nY/kwkYx3fRvtkt4Yx5nOsk/giERkoIr523fBLwL+NMTvreekLWF/QRrs72+t4CnhVrO6QnUQk0K5Hv8euO5+K9WNnmP0YinWSvg5ARH5uv7aziPiJyBys78jaFu76K1i9rSZhJSTs7Vxj19nXVFVC079bTeLMfRGRP4nIIHs9ocCPsaolj2K9hxeK1U3d1z7mk8VqmwTnfYcWYLXT3YT1uagxB+tcN8zhcSlwnt2uVZdHgV+LSGBDG7TPPxcCF9XUMTnMy8Kqjv8/e5+HYJUQa75LC4F7RSTSPhY/cXj5t0CxiNxtvz++9vEd2VA8dQXY6gdWD45vsapR8oBVWHVjnez5l2H92inG6g3xBHZdGFbd3nasX+OHgH/i0EOg1nYyqVWvCfwbWGQ/n8/pPXPuxarWyMH6wBlq9Xypta4ZWFm9xN6PJcD59jwfrMbfLHveS0Ckw2vnYmX1w1iNaqdipVbdH7V6QmCVxPZjfZF/VUdck6mjrthhu8trTQvESmJ7sBLvVk7v1XCPw3FprHfTJcAG+73bBUy3p4/FSo7HgH8ahzpW83398kv2scqyj51PAzE7vvY+4KMG3qczPge19v1P9jZP2DHfU7Nte5klnNmecjeQ4fB+VdifA8dHrD1fsE7Km7E+8weA14GB9ra+qyOuBHudg7C+G99htYkUYH13LmjOftZarjvWyf/DWtNfwvo8ltixXtzIeur8nNlx1PS+q3m8bc9z2r5gdXbYyfffvw+AAQ7zRwNLsdrl8rAasrvX93ls6cP+fBwDAuz/x2CVCmLqWHYzcAd1924Se/5P6tnOqc98remn2iTs/xPtY5GPVf17q8O8IKzEVkD9vZtexfq+H8Nqm6vzvFTfQ+yFvZ5dR7oJ642vbGx5pZRSXnZbjtpE5BKxrsCNxPp1+b4mCKWUajqvThJY3coOYxXRqrCqnJRSSjVRh6luUkop1XzeXpJQSinVCpoklFJK1cst7hgaHR1tUlJSXB2GUkp5lO++++6IMaY5F+42m1skiZSUFDIyMlwdhlJKeRQRaei2Ok6h1U1KKaXqpUlCKaVUvTRJKKWUqpdbtEkopdpeRUUF2dnZlJWVuToU1UyBgYEkJibi79+UO487lyYJpTqI7OxsQkNDSUlJoWl3RVfuwBjD0aNHyc7OpkePHu2+/Uarm0QkyR4waIuIbBaRn9nTo0TkMxHZaf+NtKeLiPxTRHaJyAYRGdHWO6GUalxZWRldunTRBOFhRIQuXbq4rATYlDaJSuCXxhoUaAxwu4ikYt0S+QtjTB+se/jfYy8/E2vs5D5YtxH+t9OjVkq1iCYIz+TK963RJGGMOWiMWWM/L8Yam6Ab1hgINYNyvABcbD+fBSwwlpVAhIg0aYhBpZRS7qVZvZtEJAUYjjWoUJwx5qA9K5fvx03txuljrmZz+tjGSinlFDk5OVx22WWuDqNO8+fPJycnx9VhtFqTk4Q9ZOci4OfGmCLHeca6lWyzbicrIjeLSIaIZOTl5TXnpUopRWVlJQkJCbz55puuDqVO3pIkmtS7yR63eRHwsjHmLXvyIRGJN8YctKuTDtvTD3D6wNyJfD9o9ynGmHnAPID09HS9X7lS7ejB9zezJaeo8QWbITUhjN9fOLDBZR566CFeeuklYmJiSEpKIi0tjfDwcObNm8fJkyfp3bs3L774IkFBQbzxxhs8+OCD+Pr6Eh4ezrJly5g/fz5vvfUWJSUlVFVV8cILL3DBBRewadMmSktLmTt3Lps2baJfv37k5OTw5JNPkp6eXmcsH3/8Mffddx9VVVVER0fzxRdfkJ+fz/XXX8+ePXsICgpi3rx5DBkyhAceeICQkBB+9atfATBo0CA++OADAGbOnMmECRP45ptv6NatG++++y4ffvghGRkZXH311XTu3JkVK1bQuXNnpx7v9tKU3k0CPAdsNcb81WHWe1iDg2P/fddh+nV2L6cxQKFDtZRSTVZWUcXmnEJXh6GcZPXq1SxatIj169fz0Ucfnbpf2+zZs1m9ejXr169nwIABPPfccwD84Q9/4JNPPmH9+vW89957p9azZs0a3nzzTZYuXXra+p966ikiIyPZsmULDz30EN999129seTl5XHTTTediueNN94A4Pe//z3Dhw9nw4YNPPLII1x33XWN7tfOnTu5/fbb2bx5MxERESxatIjLLruM9PR0Xn75ZdatW+exCQKaVpIYD1wLbBSRdfa0+4BHgYUicgOwD7jcnvc/4DysAehLgR85M2DVMVRVG257eQ1fbjvMpL4x/Oa8AfTrGurqsLxGY7/428LXX3/NrFmzCAwMJDAwkAsvvBCATZs2cf/991NQUEBJSQnTp08HYPz48cydO5fLL7+c2bNnn1rPOeecQ1RU1BnrX758OT/72c8A65f+kCFD6o1l5cqVTJo06dR1BzXrW758OYsWLQLg7LPP5ujRoxQVNVzi6tGjB8OGDQMgLS2NzMzMJhwNz9FokjDGLAfq6381tY7lDXB7K+NSHdxfPtnOl9sOc/GwBL7cdpiZ/1jGlaO6c+e0vsSEBrg6POVEc+fO5Z133mHo0KHMnz+fJUuWAPD000+zatUqPvzwQ9LS0k6VDIKDg9s9Rj8/P6qrq0/973jNQkDA959HX19fTpw40a6xtTW9d5NyO2+vzebppbv54eju/O2KYSy9awrXjU1h4eospjy2hKeW7KKsosrVYapmGj9+PO+//z5lZWWUlJScqtMvLi4mPj6eiooKXn755VPL7969m9GjR/OHP/yBmJgYsrKy6lv1qfUvXLgQgC1btrBx48Z6lx0zZgzLli1j7969AOTn5wMwceLEUzEsWbKE6OhowsLCSElJYc2aNYBV3VXzuoaEhoZSXFzc6HLuTm/LodzK2v3HuHvRRkb3iOLBiwYiIkQGd+KBiwZy7dhk/u9/2/jzx9t5a80B3rx1LBFBnVwdsmqikSNHctFFFzFkyBDi4uIYPHgw4eHhPPTQQ4wePZqYmBhGjx596sR61113sXPnTowxTJ06laFDh7Ju3bp613/bbbcxZ84cUlNT6d+/PwMHDiQ8PLzOZWNiYpg3bx6zZ8+murqa2NhYPvvsMx544AGuv/56hgwZQlBQEC+8YF0Kdumll7JgwQIGDhzI6NGj6du3b6P7O3fuXG699VaPb7gWq3bItdLT040OOqRyC8u46InldPLz4b07JhAVXHcCWLztMLe8+B3pKZG8cP0o/H21QNwUW7duZcCAAS6NoaSkhJCQEEpLS5k0aRLz5s1jxAjn3LmnqqqKiooKAgMD2b17N9OmTWP79u106uQdPyTqev9E5DtjTN3dt5xESxLKLZRVVHHzixkcL69kwQ3j6k0QAFP6x/LI7MH86o31PPDeZv548SC93YSHuPnmm9myZQtlZWXMmTPHaQkCoLS0lClTplBRUYExhqeeesprEoQraZLoAPbklZB//CSFJyooKqugsLSCorJKukcFcfFw118Mb4zh129uYOOBQp65Jo3+XcMafc1laYnsPFzMM0v30DculDnjUto+UNVqr7zySputOzQ0tM5hkEePHk15eflp01588UUGDx7cZrF4E00SXm7est088r9tdc4TgaFJEfSIbv/eIo6e+HIX763P4a7p/Th3YNcmv+7X0/uz+/Bx/vDBFnpEBzOpb5uOB6881KpVq1wdgkfTJOHFCk9U8MSXuxjfuwu3TOpFeGd/wjr7E97Zn5OV1Uz6y2Ke/WoPj1ziul9U7647wOOf7eCS4d24bXKvZr3W10f4+5XDuOzf33D7K2t4+7bx9I4NaaNIleqYtMXPi/33670UlVVy78wBTOobc6rUEBXcia7hgVw6IpE3v8smr7i88ZW1gW/35nPXGxsY1SOKRy8d3KJ2hZAAP/4zJ50APx9ufGE1BaUn2yBSpTouTRJeqrC0gue+2sv0gXEM6lZ3N8CbJvagoqqaBSsy2zc4YO+R49z8YgaJkZ2Zd20aAX6+LV5XYmQQz1ybRk5BGT9+aQ0nK6sbf5FSqkk0SXip55bvobi8kp9Pq78/d8+YEKandmXBin0cL69st9jyj5/kR//9Fh8R/vujkU651iEt2SqNrNhzlN++swl36NqtlDfQJOGFCkpP8vzXmZw3uCsD4hvuKXTLWT0pPFHB66sbvprVWcoqqrh5QQY5hWU8e10ayV2c12g+e0QiPzm7N69nZDFv2R6nrVe5r9aMJ/HGG28wYMAApkyZ4uSoWmbJkiV88803rg7jDNpw7YWe/WoPx09W8rOpjV8VOrx7JKN6RPHc8r1cOza5TS9Mq+nqmrHvGE/8cDhpyWfepK217pzWlz1HjvPox9tI7hLMjEFN7y3VYTxQd/Wj89bfPnfube14Es899xzPPvssEyZMaPL2/Pza7pS5ZMkSQkJCGDduXJttoyW0JOFl8o+fZP7XmZw3OL7Jd0299ayeHCg4wYcb2vaO7v/84vuurhcMSWiTbfj4CI//YChDEyP4+etr2Zittxp3Jw899BD9+vVjwoQJXHXVVTz22GM8++yzjBw5kqFDh3LppZdSWloKWL/0Bw0axNChQ5k0aRJgDeRz0UUXcfbZZzN16lQyMzMZNGgQYF1Md/nll5Oamsoll1zC6NGj67xuAqzbkC9fvpwbbriBu+66i7KyMn70ox8xePBghg8fzuLFi+vc3vHjx7n++usZNWoUw4cP5913rRESqqqq+NWvfnXq7rP/+te/Tm1n5MiRDBo0iJtvvvlUNeg///lPUlNTGTJkCFdeeSWZmZk8/fTT/O1vf2PYsGF89dVXbfcmNJOWJLzMs1/tobSiip9P7dPk10zuG0vfuBCeXrqbWcMS2uTq5Y83HeRvn+9gdgu6ujZXoL8vz16XzsVPfs0NL6zm3TvGEx/umffN8SaO40lUVFQwYsQI0tLSmD17NjfddBMA999/P8899xw/+clPTo0n0a1bNwoKCk6tZ82aNWzYsIGoqKjTbsvtOJ7Epk2bTt2+uy6/+93v+PLLL3nsscdIT0/n8ccfR0TYuHEj27Zt49xzz2XHjh1nbO++++7j7LPP5vnnn6egoIBRo0Yxbdo0FixYQGZmJuvWrcPPz+/UDQPvuOMOfve73wFw7bXX8sEHH3DhhRfy6KOPsnfvXgICAigoKCAiIoJbb731tIGN3IWWJLzI0ZJyXvgmkwuHJNAnruljL/j4CDdN7Mm23GKW7Tzi9Li25BRx5+vrGd49gkdmt6yra3PFhAbw/NyRlJ6s4vr5Ge3aMK/q5jieRGho6GnjSUycOJHBgwfz8ssvs3nzZuD78SSeffZZqqq+v+tvQ+NJXHnllUDj40nU9dprrrkGgP79+5OcnHwqSThu79NPP+XRRx9l2LBhTJ48mbKyMvbv38/nn3/OLbfccqo6qmb5xYsXM3r0aAYPHsyXX355at+GDBnC1VdfzUsvvdSmVVjO4N7RqWaZt2wPZRVV/LQZpYgas4Z147FPt/PM0t2c5cQrl4+UlHPTggzCO/vzzDVpBPq3vKtrc/XrGsqTV4/g+vmruf+dTfztimHttm231k5tBk3l7uNJOG7PGMOiRYvo169fo68rKyvjtttuIyMjg6SkJB544IFT41B8+OGHLFu2jPfff5+HH364wduau5qWJLzEkZJyFqzYx6xh3Vp01XEnPx9umNCDb3YfZUN2gVNiOllZzY9f+o4jJeU8e106sWGBTllvc5zVN4Yfn9WLt9ceYM3+Y+2+ffU9dxpPojbHcSR27NjB/v3760wE06dP51//+teptoW1a9cCVmnjmWeeobLSKrHm5+efSgjR0dGUlJScamCvrq4mKyuLKVOm8Kc//YnCwkJKSkrcdvwJTRJe4rVv93Oiooo7zu7d4nVcNao7oQF+/HvJ7lbHY4zht+9sYnXmMR77wVAGJ7Zxj5oG/HhyL2JCA/jjB1v0+gkXchxPYubMmWeMJzF+/Hj69+9/avm77rqLwYMHM2jQIMaNG8fQoUMbXP9tt91GXl4eqamp3H///Q2OJ1HXa6urqxk8eDBXXHEF8+fPP23EuRq//e1vqaioYMiQIQwcOJDf/va3ANx44410796dIUOGMHToUF555RUiIiK46aabGDRoENOnT2fkyJGA1ch9zTXXnGok/+lPf0pERAQXXnghb7/9tts1XGOMcfkjLS3NqJarqqo2E//0pbnymRWtXtfjn2wzyXd/YNbuP9aq9Ty/fI9JvvsD85ePt7U6Jmd4/dv9JvnuD8x76w64OhSX2bJli6tDMMXFxcYYY44fP27S0tLMd99957R1V1ZWmhMnThhjjNm1a5dJSUkx5eXlTlu/q9X1/gEZpo3Pz1qS8AKr9uazP7+Uy0cmtnpdN5/Vi+iQTjzy4dYW/+redKCQP364lXNS4/jFOY1fq9EeLk1LZEB8GI9+tE2HPnWhm2++mWHDhjFixAguvfRSp48nMWHCBIYOHcoll1yi40k4iTZce4GFGVmEBvoxc1B8q9cVEuDHnef05Tdvb+LTLYeY3oxbdwNUVlVz96INRAV34rEfDMXHxz0GA/L1Ee4/fwBX/2cV//06kx+3cTdcVTcdT8LzaJLwcEVlFfxv40EuS0t0Ws+hK9KT+O/XmTz60TbO7h/brKuwn/96L5tzinjq6hGEd/Z3SjzOMr53NNMGxPLk4l1clpZITOiZdc7ezhjT4Ubx84bxJFpaqncGrW7ycO+ty6G8sporRiY5bZ1+vj7cd15/9h45ziur9jf5dfuPlvLXz3YwbUAcM930dhj3njeAsooq/vb5DleH0u4CAwM5evSoNt57GGMMR48eJTCw/XsHgpYkPN4bGVn07xrK4HpuB95SU/rFMq5XF/7++Q4uGdGNsMCGSwXGGH7zzkb8fHx46OKBbvtrtVdMCNeMSWbBikzmjE1p8q1LvEFiYiLZ2dnk5eW5OhTVTIGBgSQmtr7NsSU0SXiwbblFrM8u5HcXpDr9pCwi3HfeAC58YjlPLd7NPTP7N7j822sP8NXOI/xh1kC3vwXGz6b24e21B/jjh1tYcP0ot01ozubv70+PHj1cHYbyMB22umnvkeMcKiqjutpzi94LV2fj7ytcPLxbm6x/ULdwLhnejee/3kv2sdJ6lztaUs5DH2xhRPcIrhmd3CaxOFNkcCd+OrUPX+08wpId+qtaqYZ0yJLE6sx8fvD0CsC60jgxojPdIjuTFBVEz+hgZg3r5vaNmuWVVby9NptzU7sSFdx23fx+dW4/PtxwkMc+2c7frxxe5zJ//HArJeWVPHrpELfpzdSYa+0qp8c+2c7kvjEdpjShVHN1yJLE00t2ExXciYdmDeRH41IYEB9G4YkKPt6Uyx8/3Mr4P33JvW9tZE9eiatDrdcXWw9zrLSCH6S3bT1lQkRnbpzYg3fW5fD5lkPsP1rK4eIyissqqKiqZumOPN5ee4Afn9WLvs24qaCrdfLz4Sdn92FzThGfbjnk6nCUclsdriSx81AxX2w7zJ3T+nLt2JQz5u89cpxnv9rDm99l89rq/UxP7cotZ/VkePfI9g+2AQszsogPD2RiH+fdjK8+t57Vi9dXZ3Hjgrrvzd8zJpjbprT8diCucvGwBJ5cvIu/f76TcwbEeUwpSKn21OGSxLxlewj09+HasXXXnfeIDuaRSwZz57S+vPBNJgtWZPLx5lxG9Yjil+f0ZXTPLu0c8ZkOFp5g2Y48bp/SG992OLGFBvrz/k8msGZfAWUVVZyoqLL+nqziZFU1s0c47xqN9uTn68NPzu7NLxau59MtucxwwsWISnmbDpUkDhWV8c66A1w1qnuj9fgxoQH8ano/fjy5F6+tzuKZpbu5Yt5KJvaJ5q7p/RiSGOG0uN78LpvqasO01LgmtS8s+i6bagM/SHPetRGNiQ/vzPlD3LvXUktcNDSBJ760ShPnpnbV0oRStXSoJPHfrzOpqjbcOKFnk18THODHDRN6cPXo7ixYkcm/l+zmoie+ZsbArvzi3L6troc/WHiCu95cjzHg8xaM7tGF6QPjmD6o62ldSY0xFJVVkldcxsKMbMb27EL3LkGt2rayShM/ndqHn7++jo825XL+EC1NKOWowySJ4rIKXl61j5mD4lt0cg309+XmSb24alR3nlu+l/98tZdPtuTyo3E9+N2FqS2O6711ORgD865NY+OBQj7elMsD72/hgfe3MLhbOP6+wuHicvKKyymvrD71ul+e6x43zvMGFw5N4F9f7uQfX+xg5iAtTSjlqMMkide+zaK4rJKbJzW9FFGX0EB/fj6tL3PGpvDg+5t5/uu9XDOmOz1jmj/QD8A763IYlhTBuQO7cu7Arvzy3H7szivhk825LNmeh5+PkJ4cSWxYILGhAcSEBpAQ0Zk0N2tI92S+PsJPp/bhZ6+t48ONB7lwaIKrQ1LKbXSIJFFRVc3zX+9lTM8ohiZFOGWdkcGduPe8Aby7Pod31h7gF+c2Ppxhbdtzi9l6sIgHLxp42vReMSHcNrk3t032vB5DnuqCIVbbxD++2Ml5g+PbpUOAUp6gQ1wn8f76HA4WlnHLJOfeHjouLJDxvaJ5e92BFt007Z11B/D1Ea0HdwO+PsLPpvVh1+ESPtiQ4+pwlHIbXp8kjDHMW7aHvnEhTO7n/GsKLhnejaz8E80eP7m62vDu2gNM6hNNdIh7X93dUZw3KJ5+caH884udVHnw7VqUciavTxJLd+SxLbeYmyb2bJNbL0wf1JVAfx/eWnOgWa9bnZlPTmFZm913STWfj12a2J13XEsTStkaTRIi8ryIHBaRTQ7ThonIShFZJyIZIjLKni4i8k8R2SUiG0TEeWMTttC8ZXuICwtg1rC2ORmHBPhxbmpXPthwkJMOvY8a8866AwR18uWc1Lg2iUu1zIyBXekRHcz8bzJdHYpSbqEpJYn5wIxa0/4MPGiMGQb8zv4fYCbQx37cDPzbKVG2UFZ+Kd/sPsp1Y1Po5Nd2haZLRnSj8EQFS7YfbtLy5ZVVfLjhINMHdiWoU4foO+AxfHyEa8cks3Z/ARuyC1wdjlIu1+iZ0xizDMivPRkIs5+HAzVl81nAAmNZCUSIiMtaZZfttG4DPX1g2/5an9g7mi7BnXh7bdOqnBZvy6OorJJZw7SrpTu6LD2RoE6+vPDNPleHopTLtfTn9c+Bv4hIFvAYcK89vRuQ5bBctj3tDCJys11VldFWI2Ut25FHQnggvVp4DUNT+fn6cOHQBL7YepjCExWNLv/uugNEh3RiQu/oNo1LtUxYoD+zR3Tj/Q05HC0pd3U4SrlUS5PEj4E7jTFJwJ3Ac81dgTFmnjEm3RiTHhPj/F5HFVXVfLPrKJPaaayA2SO6cbKqmo82HmxwucITFXyx7TAXDEnAz9fr+w14rDljUzhZWc3rGVmNL6yUF2vpWWoO8Jb9/A1glP38AOB417lEe1q7W5dVQHF5JZP6tv2ttAEGdwunZ0wwbzVS5fTxJquB+xLt1eTW+sSFMq5XF15asY/KqqZ3SFDK27Q0SeQAZ9nPzwZ22s/fA66zezmNAQqNMQ3/tG4jy3bk4SMwvlf7VOmICJcM68a3e/MbHOrznbU59IgOZkhieLvEpVruurEp5BSW8fnWpnVIUMobNaUL7KvACqCfiGSLyA3ATcDjIrIeeASrJxPA/4A9wC7gWeC2Nom6CZbtyGNYUgThQf7tts2aax7eXVd3H/uDhSdYufcos4Yl6HCZHmDagFi6RXRmwYpMV4eilMs02v/SGHNVPbPS6ljWALe3NqjWyj9+kg0HCvnZ1D7tut2kqCBGpkTy9toD3Da51xmJoOaOrxe30TUbyrn8fH24ekx3/vzxdnYcKvao4VmVchavbDldvusIxtBu7RGOLh7ejV2HS9icUwTAnrwS/vPVHq7+z0oe+3Q7w7tHkBId3O5xqZa5cmR3Ovn5aGlCdVheeSXXsh15hHf2Z6gTR49rqgsGJ/Dge1u4560NlJRVknnUap/oExvC9eN71DtsqnJPUcGduGhoAm+tOcCvZ/QnLLD9qi+VcgdelySMMXy1M48JvaNdcrvn8CB/zhvclY825TKuVxdumNCDyf1iSYrSUeQ81ZyxKbz5XTZvZmRz/YQerg5HqXbldUli+6FiDhWVM6mv6y5U++vlw/jzZaZNbwWi2s/gxHBGdI/gxZX7mDsuRUeuUx2K153Flu2wrt52RXtEDR8f0QThZeaMS2HvkeMs33XE1aEo1a687ky2bMcR+sSGEB/e2dWhKC8yY1BXIoP8eX21XoGtOhavShInTlbxbWa+S0sRyjsF+Ply8fBufLoll/zjJ10djlLtxquSxMq9RzlZWa1JQrWJK0YmUVFlmny3X6W8gVcliWU78gjw82F0jyhXh6K8UP+uYQxNimDh6qwWjWmulCfyqiTx1c4jjOoRRaC/r6tDUV7qivQkth8qZn12oatDUapdeE2SyCk4wa7DJZylVU2qDV04NJ7O/r7agK06DK9JEu7Q9VV5v9BAf84bHM/763MoPVnp6nCUanPekyR25tE1LJA+sW07Cp1SV4xMoqS8kg83uOQu+Eq1K69IEtXVhq93HWVin2i9BbdqcyNTIukZHcxCHbVOdQBekSQyjx6n8EQF6SmRrg5FdQAiwuUjk1ideYzdeSWuDkepNuUVSWJ9dgEAQ5MiXBqH6jhmj+iGr49oaUJ5Pe9IElmFBHXypU+sDgqj2kdsaCBn949l0XcHqNAxsJUX844kkV3AoG7hLrk1uOq4rkhP4khJOYu36RjYynt5fJI4WVnN5pwihmlVk2pnk/vFEBsaoNdMKK/m8Ulie24xJyurXTIKnerY/Hx9uDQtkcXbD5NXXO7qcJRqEx6fJNadarQOd20gqkM6f3A81Qa+1nEmlJfy+CSxPquA6JBOdIvQ8SNU+xsQH0ZYoB8rdh91dShKtQmvSBJDEyP0IjrlEr4+wpieXVixR5OE8k4enSSKyyrYlVei10colxrbqwv780vJPlbq6lCUcjqPThIbDxRiDAxJ1PYI5TrjekUDaJWT8koenSQ22Pf0155NypX6xoXQJbiTJgnllTw6SazPKiC5SxCRwZ1cHYrqwESEMb2sdgkdsU55G49PElqKUO5gbM8uHCwsI/Ootkso7+KxSeJwURk5hWXaaK3cwrheXQBtl1Dex2OTRM0Yw8P0IjrlBnpEBxMXFsA3u/WiOuVdPDdJZBXg6yMMTNAkoVxPRBjXK5qV2i6hvIznJonsAvp3DSXQ39fVoSgFWO0SR0pOsvOwDkSkvIdHJonqamM1Wmt7hHIjY7VdQnkhj0wSmUePU1RWyTDt2aTcSFJUEImRnbVdQnkVj0wSOlypcldje3Zh5Z58qqu1XUJ5B89MEvZwpb1jQ1wdilKnGde7C4UnKthysMjVoSjlFJ6ZJHS4UuWmxva07uO0Uu8Kq7yExyUJHa5UubOu4YH0jA7WxmvlNRpNEiLyvIgcFpFNtab/RES2ichmEfmzw/R7RWSXiGwXkenODliHK1XubkyvLqzam09lVbWrQ1Gq1ZpSkpgPzHCcICJTgFnAUGPMQOAxe3oqcCUw0H7NUyLi1AsZdLhS5e7G9epCSXklm3K0XUJ5vkaThDFmGZBfa/KPgUeNMeX2Moft6bOA14wx5caYvcAuYJQT49XhSpXbG9PTul5Cu8Iqb9DSNom+wEQRWSUiS0VkpD29G5DlsFy2Pe0MInKziGSISEZeXl6TNmqM4du9+QxL0uFKlfuKDgmgX1yotksor9DSJOEHRAFjgLuAhdLMs7YxZp4xJt0Ykx4TE9Ok12zLLWZ/filTB8Q1O2Cl2tPYXl1YnZlPhbZLKA/X0iSRDbxlLN8C1UA0cABIclgu0Z7mFB9vykUEzknVJKHc2/DuEZRVVLM7T+/jpDxbS5PEO8AUABHpC3QCjgDvAVeKSICI9AD6AN86IU4APtmcy8iUKKJDApy1SqXaRGp8GABbtPFaebimdIF9FVgB9BORbBG5AXge6Gl3i30NmGOXKjYDC4EtwMfA7caYKmcEuvfIcbblFjNjYFdnrE6pNtUjOpgAPx9NEsrj+TW2gDHmqnpmXVPP8g8DD7cmqLp8sjkXgOmDNEko9+fn60P/rqFszdUkoTybx1xx/fGmXIYkhmvXV+UxUhPC2JJTpIMQKY/mEUniYOEJ1mUVMF2rmpQHGRAfxrHSCnKLylwdilIt5hFJ4tPNhwCYoVVNyoNo47XyBh6RJD7elEuf2BB6xeitwZXn6K9JQnkBt08S+cdPsmrvUS1FKI8TEuBHSpcgHVtCeTS3TxKfbzlEtUHbI5RHSk0IY6smCeXB3D5JfLw5l8TIzgxMCHN1KEo1W2p8GJlHSykpr3R1KEq1iFsnieKyCpbvPMKMgV31hn7KIw2w2yW2aWlCeSi3ThKLt+dxsqpa2yOUx0q1S8DaLqE8lVsniU825RITGsCI7pGuDkWpFukaFkhkkL/2cFIey22TRFlFFYu3H+bc1Dh8fLSqSXkmEbGuvNaShPJQbpsklu3Io/RklVY1KY+XGh/G9txiHfNaeSS3TRIfb8olLNDv1FCQSnmq1IQwyiur2XvkuKtDUarZ3DJJnDhZxSebc5kxqCv+vm4ZolJNVtPDSauclCdyyzPw51sPcfxkFRcPq3N4bKU8Sq+YEDr56tgSyjO5ZZJ4d90BuoYFMlqrmpQX8Pf1oW/XEC1JKI/kdkki//hJlmzP46JhCfhqryblJVLjdWwJ5ZncLkl8uPEgldWGWcMSXB2KUk6TGh/G0eMnOVxc7upQlGoWt0sS7649QJ/YkFP34lfKG6QmhAPaeK08j1sliaz8UjL2HePi4d30Xk3Kq/SPDwV0bAnledwqSby3PgeAi4ZqVZPyLmGB/iRFddaShPI4bpMkjDG8s/YAI1MiSYoKcnU4SjldanwYW7UkoTyM2ySJLQeL2Hm4hFl6bYTyUqnx4ew9epzjOraE8iBukyTeWXsAPx/h/MHxrg5FqTaRmhCGMbAtt9jVoSjVZG6TJN5bn8PkfjFEBndydShKtYmasSV0OFPlSdwiSZSUV3KoqFyrmpRXSwgPJLyzvzZeK4/iFkmioLSC4E6+TBsQ5+pQlGozIsKA+FDtBqs8ilskicITFcwYFE/nTr6uDkWpNjUoIZwtB4s4WaljSyjP4BZJotoYLh6u10Yo75eWHMnJymo25xS6OhSlmsQtkoSfjzCuV7Srw1CqzaUlW+O1f7fvmIsjUapp3CJJ9IkL1Tu+qg4hNiyQpKjOmiSUx3CLJOGnCUJ1IGndI8nYd0xvG648glskCaU6krTkSPKKy8k+dsLVoSjVKE0SSrWztOQoQNsllGfQJKFUO+vXNZSQAD9NEsojaJJQqp35+gjDu0eQoUlCeQBNEkq5wIjukWzPLaK4rMLVoSjVoEaThIg8LyKHRWRTHfN+KSJGRKLt/0VE/ikiu0Rkg4iMaIuglfJ0acmRVBtYl1Xg6lCUalBTShLzgRm1J4pIEnAusN9h8kygj/24Gfh360NUyvsM7x6BiDZeK/fXaJIwxiwD8uuY9Tfg14BjZ+9ZwAJjWQlEiIgOEKFULaGB/vSLC9Ukodxei9okRGQWcMAYs77WrG5AlsP/2fY0pVQt6SmRrN1fQFW1XlSn3Fezk4SIBAH3Ab9rzYZF5GYRyRCRjLy8vNasSimPlJYcSUl5JTsO6Uh1yn21pCTRC+gBrBeRTCARWCMiXYEDQJLDson2tDMYY+YZY9KNMekxMTEtCEMpz5bW3bqoTrvCKnfW7CRhjNlojIk1xqQYY1KwqpRGGGNygfeA6+xeTmOAQmPMQeeGrJR3SIrqTExoAGs0SSg31pQusK8CK4B+IpItIjc0sPj/gD3ALuBZ4DanRKmUFxIR0rpHauO1cmt+jS1gjLmqkfkpDs8NcHvrw1KqY0hPieTjzbkcLi4jNjTQ1eEodQa94lopFxphD0KkVU7KXWmSUMqFBiaE0cnPh4xMTRLKPWmSUMqFAvx8GZoYznf7NUko96RJQikXG5EcyaYDhZRVVLk6FKXOoElCKRdLT46iosqw8UChq0NR6gyaJJRysRHdIwC92Z9yT5oklHKxLiEB9IgOZvXeuu6jqZRraZJQyg2c3T+WpTvyOFh4wtWhKHUaTRJKuYG541KoNob532S6OhSlTqNJQik3kBQVxMzB8byyaj8l5ZWuDkepUzRJKOUmbprYk+KySl5fndX4wkq1E00SSrmJYUkRjEyJ5Pnle6msqnZ1OEoBmiSUcis3TuzJgYITfLw519WhKAVoklDKrUwbEEdKlyCeXbYH66bKSrmWJgml3Iivj3DDxJ6szy5ktd70T7kBTRJKuZnLRiQSGeTPs1/tcXUoSmmSUMrddO7kyzVjkvl86yH2Hjnu6nBUB6dJQik3dO3YZPx9fHhuuZYmlGtpklDKDcWGBnLx8ATeyMgm//hJV4ejOjBNEkq5qRsn9qS8spqXVu5zdSiqA9MkoZSb6hsXyqS+Mby4ch8nK/XiOuUamiSUcmM/Gp9CXnE5H2066OpQVAelSUIpN3ZWnxh6RAfz368zXR2K6qA0SSjlxnx8hDljk1mXVcDa/XpxnWp/miSUcnOXpiUSEuDHCzrWhHIBTRJKubnQQH8uS0vkw40HOVxU5upwVAejSUIpDzB3XAqV1YaXV+13dSiqg9EkoZQHSIkOZkq/WF5etZ/yyipXh6M6EE0SSnmIueNSOFJSzv82andY1X40SSjlISb2iaZXjNUdVseaUO1Fk4RSHkJEmDsuhQ3ZhazNKnB1OKqD0CShlAeZPSKR0AA/5uvFdaqdaJJQyoMEB/hx+cgk/rfxIIe0O6xqB5oklPIw141NpsoYXl+d5epQVAegSUIpD5PcJZhhSRF8ue2wq0NRHYAmCaU80KQ+MWzILuCYDkik2pgmCaU80Fn9Yqg2sHzXEVeHorycJgmlPNDQxAjCO/uzbEeeq0NRXq7RJCEiz4vIYRHZ5DDtLyKyTUQ2iMjbIhLhMO9eEdklIttFZHobxa1Uh+brI0zoE82ynXl6YZ1qU00pScwHZtSa9hkwyBgzBNgB3AsgIqnAlcBA+zVPiYiv06JVSp1yVp8YDhWVs/1QsatDUV6s0SRhjFkG5Nea9qkxptL+dyWQaD+fBbxmjCk3xuwFdgGjnBivUso2qW8MAEu3a5WTajvOaJO4HvjIft4NcOy8nW1PO4OI3CwiGSKSkZenH3KlmqtreCD9u4ayVNslVBtqVZIQkd8AlcDLzX2tMWaeMSbdGJMeExPTmjCU6rAm9Y0hI/MYx8srG19YqRZocZIQkbnABcDV5vuWswNAksNiifY0pVQbOKtvDCerqlm556irQ1FeqkVJQkRmAL8GLjLGlDrMeg+4UkQCRKQH0Af4tvVhKqXqkp4SSWd/X+0Kq9qMX2MLiMirwGQgWkSygd9j9WYKAD4TEYCVxphbjTGbRWQhsAWrGup2Y4wOo6VUGwnw82Vsry7aLqHaTKNJwhhzVR2Tn2tg+YeBh1sTlFKq6Sb1iebLbYfZd/Q4yV2CXR2O8jJ6xbVSHu6sfrEAWuWk2oQmCaU8XEqXIJKiOmuVk2oTmiSU8nAiwll9Y/hm91FOVla7OhzlZTRJKOUFJvWJofRkFRn78htfWKlm0CShlBcY1zsaPx9h2Q69dbhyLk0SSnmBkAA/0lMitV1COZ0mCaW8xKS+MWw9WMThojJXh6K8iCYJpbzEWfZdYb/Qsa+VE2mSUMpLpMaH0Sc2hFe/3e/qUJQX0SShlJcQEa4dm8yG7ELWZRW4OhzlJTRJKOVFLhnejeBOvry4Yp+rQ1FeQpOEUl4kNNCf2SMSeX9DDvnHT7o6HOUFNEko5WWuHZvMycpqFmZkNb6wUo3QJKGUl+kbF8roHlG8vGofVdWm8Rco1QBNEkp5oevGppCVf4KlO7Q7rGodTRJKeaFzB8YRGxrAAm3AVq2kSUIpL+Tv68NVo7qzdEce+44ed3U4yoNpklDKS/1wdHd8RHh5lV5cp1pOk4RSXiouLJDpA+NYmJFFWYUONa9aRpOEUl7s2jEpFJRW8P76HFeHojyUJgmlvNiYnlH0iQ3hxZXagK1aRpOEUl7M8X5OG7MLXR2O8kCaJJTycrOGdsPXR/h480FXh6I8kCYJpbxceJA/6cmRfLFVL6xTzadJQqkO4JzUOLblFpOVX+rqUJSH0SShVAcwdUAcAF9sPeTiSJSn0SShVAfQIzqYnjHBOrSpajZNEkp1EOcMiGPlnqMUl1W4OhTlQTRJKNVBTB0QR0WVYdmOI64ORXkQTRJKdRAjukcQEeSv7RKqWTRJKNVB+Pn6cHa/WBZvP6yDEakm0yShVAcydUAcx0orWLP/mKtDUR5Ck4RSHcikvtH4+wqfb9EqJ9U0miSU6kBCA/0Z3aMLn2u7hGoiTRJKdTDTBsSyO+84e4/oiHWqcZoklOpg9Opr1RyaJJTqYJKigugXF6pVTqpJ/BpbQESeBy4ADhtjBtnTooDXgRQgE7jcGHNMRAT4B3AeUArMNcasaZvQlVItNS01lqeX7qGwtILwIH9Xh9OhbM8t5sONB+kdG8LwpAgSIztjnTpPV1lVzc7DJWzILiC3sNwFkVoaTRLAfOAJYIHDtHuAL4wxj4rIPfb/dwMzgT72YzTwb/uvUsqNTB0Qx5OLd7Nkx2FmDevm6nA6hLKKKp5avIt/L91NRdX316lEh3RiWFIkw7tHEBsawOacIjYeKGRzTiFlFdUujNjSaJIwxiwTkZRak2cBk+3nLwBLsJLELGCBMcYAK0UkQkTijTE62olSbmRYYgTRIZ34fKsmifbw7d587nlrA3vyjjN7eDfuOa8/h4vKWZtVwLr9BazNOnaq+q+zvy+DuoXxw1HJDEkMZ0hiON2jgvCpo7Th+6e2j70pJYm6xDmc+HOBOPt5NyDLYblse5omCaXciI+PcHb/WD7alEtFVTX+vto82RaKyip49KNtvLJqP4mRnVlw/Sgm9Y0BIDY0kEHdwrl2TDIAhaUV5JWU0yM6GF+fMxOCq7Q0SZxijDEi0uxr/EXkZuBmgO7du7c2DKVUM53VN5aFGdlsySliaFKEq8PxKjsOFfP++hxeX53FkZJybpzQg1+c25egTvWfcsOD/N2yfailSeJQTTWSiMQDNTepPwAkOSyXaE87gzFmHjAPID09XW8ko1Q7G9wtHIAtBzVJOEPmkeN8sCGH99cfZPuhYnwExvWK5tnr0j36+LY0SbwHzAEetf++6zD9DhF5DavBulDbI5RyT4mRnQkN8GNLTpGrQ/Fo67IKePD9zazdXwBAenIkD140kJmDuxIbGuja4JygKV1gX8VqpI4WkWzg91jJYaGI3ADsAy63F/8fVvfXXVhdYH/UBjErpZzAx0cYEB/GloOaJFrixMkq/vrZdp5bvpfY0EDuO68/5w9JoFtEZ1eH5lRN6d10VT2zptaxrAFub21QSqn2kZoQxsKMLKqrDT5u1Fjq7lbuOco9izaQebSUH47uzr0z+xMa6H7tCc7Q6oZrpZTnSo0Po/RkFfvyS+kRHezqcNxesd1b6eVV++keFcQrN41mXK9oV4fVpjRJKNWBpSaEAbD1YJEmiUYUlVVw4b+Wk5Vfyo0TevDLc/vRuZOvq8Nqc5oklOrAeseG4OcjbMkp4rzB8a4Ox6396aNtZOWX8tINoxnX27tLD470ChqlOrBAf196x4Zo43Ujvt2bz8ur9nP9+B4dKkGAJgmlOrzU+DDtBtuAsooq7nlrA4mRnfnFuX1dHU670yShVAeXmhBGblEZR0tcd6dRd/bEl7vYk3ecRy4Z3OAV095Kk4RSHVxqfE3jdbGLI3E/Ww8W8fTS3Vw6IvHUPZc6Gk0SSnVwA+wkseVgoYsjcS9V1YZ7Fm0gvLM/958/wNXhuIwmCaU6uMjgTiSEB2q7RC3//Xov67ML+f1FA4kM7uTqcFxGk4RSitQEvT2Ho6z8Uh7/dAdT+8dy4ZCO3TVYk4RSitT4MHbnHaesosrVobjc3iPH+cmra/EReOjiQXUOLeoWTpa2y2Y6XlO9UuoMA+LDqKo27DhUzJDECFeH4xLHyyt5YvEunvtqL538fPjTZUNIcMeb9VVVwNqXYGk7DEuHJgmlFN/fnmNLTlGHSxLGGN5bn8Mj/9vKoaJyLh2RyN0z+hEb5ma3+TYGtrwDX/4Rju5qt81qklBKkRQZREiAX4drl9hxqJj7397Et5n5DO4WzlNXp5GWHOnqsM60ezF8/gAcXNfum9YkoZSyx5YI7VA9nLLyS7nimRWICI/OHszl6Unud7v0Q1vgk3thzxKXhaBJQikFWI3Xb36X3SHGligpr+TGFzKoNvDObePc7w64ZYWw5FFY9QyYOjoTiA8MvwZ4os1D0d5NSinAapc4frKK/fnt02vGVaqrDXe+vo5deSU8+cMR7pUgjIH1r8G/0mHlU3UniAEXwW2r4KJ/tUtIWpJQSgGQGh8OwJaDRaS404nTyR7/bDufbTnEAxemMqGPG93RNXcj/O8u2L+i7vkpE2Hag5CY1q5haZJQSgHQJy4EXy8fW+LddQd4cvFurhqVxJxxKa4Ox1JRBksfha//WXfJIbw7zPg/6H8+uOCaDU0SSinAHlsixnvHllifVcCv39zAqB5RPHiRm1wkl7Ua3r0djmw/c55vAIz/GUy4EzoFtX9sNk0SSqlTUhPCWLH7qKvDcLpDRWXctCCDmNAA/n31CDr5ubg5tuIELH4YVjwJpvrM+X2mw8xHIapn+8dWiyYJpdQpqfFhvL32AEdLyukSEuDqcJzm9+9upqS8kkXXj3P9fu1faZUe6rogLjQBzn8c+p/X/nHVQ3s3KaVOqbny2pvGltiTV8InW3K5fnyPU7dFd4mqSvjyYXh+Rt0JYvi1cPtKt0oQoCUJpZQDx7El3KrnTys8+9Ve/H19XNtQXZAFi26ErJVnzgtLhIv+Ab2ntX9cTaBJQil1SlRwJ+K9aGyJvOJyFq3J5tIRicSEuqiaact78N4d1gVytaX9CM75AwS6sITTCE0SSqnTpMZ7z9gSC1ZkUlFVzU0Te7T/xitOwCe/gYznzpwXEgeXPA29zm7/uJpJk4RS6jSpCWEs2ZFHSXklIQGee4ooPVnJiyv3cc6AOHrGhLTvxo/sgoXXweHNZ87rfQ5c/G8I8Ywxs7XhWil1mnG9oqmqNny964irQ2mVhauzKCit4Jaz2rkb6dYP4NkpZyYIH38492H44UKPSRCgSUIpVUt6SiShAX4s2X7Y1aG0WGVVNf9Zvpe05EjSkqPaZ6NVldbtvF+/GsprVddF9YQbPoVxd4CPZ512PbcsqZRqE/6+PkzsG83ibXkYY9zjyuRm+t+mXLKPneB3F6S2zwZL8mDR9bB32ZnzBl0KF/4DAkLbJxYn86yUppRqF5P7xZJbVOaR10sYY5i3bDc9Y4KZNiCu7TeYnQHzzjozQfj4wcw/w6XPeWyCAE0SSqk6TO5n1Zkv9sAqpxW7j7LpQBE3TezZtuNiGAMZz1sXxxUdOH1eSFeY+yGMvsUlN+VzJk0SSqkzxIYGMrhbOIu3eV6SeGbZHqJDArhkeLe220hFmXXtwwd3QnXF6fOSx8Mty6D7mLbbfjvSJKGUqtOUfjGs2X+MgtKTrg6lybblFrF0Rx5zxyUT6O/bNhspyIL/zoC1L505b+wdcN27ENoO1VztRJOEUqpOU/rHUm1g6Y48V4fSZM8v30tnf1+uGZPcNhvYs8Rqf8hZe/p0/2C47L8w/WHw9W+bbbuIJgmlVJ2GJEYQFdyJJds9I0kUlJ7k3XU5XDKiGxFBnZy78upqWP43ePESKK11K/WoXnDTFzBotnO36Sa0C6xSqk6+PsJZfWNYsv0wVdUG37ZsBHaChRlZlFdWc91YJ5ciSvPh7Vth5ydnzut3nnV7jcBw527TjWhJQilVryn9YzlWWsH67AJXh9KgqmrDiyv3MapHFP27OvFmeVmr4emJdSQIgbPvhyte9uoEAa1MEiJyp4hsFpFNIvKqiASKSA8RWSUiu0TkdRFxcrlPKdVeJvWJxkdw+15OS3ccJiv/BHPGpjhnhcZYo8b9dwYUZZ8+LzACrn4DJt3lcVdPt0SL91BEugE/BdKNMYMAX+BK4E/A34wxvYFjwA3OCFQp1f4igjqRlhzp9tdLvPDNPuLCAjh3oBN6FZ04Bq9fA5/cB9WVp89LHAm3Loc+57R+Ox6itWnQD+gsIn5AEHAQOBt4057/AnBxK7ehlHKhyf1i2XSgiMNFZa4OpU57jxxn6Y48fjgqGX/fVp7SMr+2qpe2fXDmvLF3wNz/QURS67bhYVp8RI0xB4DHgP1YyaEQ+A4oMMbUpN9soM4rWkTkZhHJEJGMvDzP6D2hVEc0pV8sgNv2cnpp5T78fYWrRrfi5F1VAV/8AeafD4VZp88LCLfaHqY/DH4dr/a8NdVNkcAsoAeQAAQDM5r6emPMPGNMujEmPSbGc26bq1RHMyA+lK5hgW5Z5VR6spKFGVnMGBRPbGhgy1ZydDc8dy589ThgTp+XMBxuXQYDLmh1rJ6qNV1gpwF7jTF5ACLyFjAeiBARP7s0kQgcaGAdSik3JyJM6R/D++sPcrKymk5+7tNY+87aHIrLKpnTkm6vxlhXTX90N1QcP3P+2Dtg6u/Az0XDnrqJ1rzb+4ExIhIk1r2EpwJbgMXAZfYyc4B3WxeiUsrVpvSLpaS8kox9+a4O5RRjDAtWZJIaH0ZacmTzXpy/F17+gXX/pdoJIqQrXPuOXb3UsRMEtK5NYhVWA/UaYKO9rnnA3cAvRGQX0AWoY4BXpZQnGd87Gn9fcauusKszj7Ett5jrxiY3fcyLynJY+hd4agzs+uzM+f0vgNtWQK8pzg3Wg7XqimtjzO+B39eavAcY1Zr1KqXcS3CAH6N7dOHjzbn88tx+bXfzvGZ4YUUmYYF+zBrWxLu97lkKH/4Sju48c55/EMx4FEZc5/G39nY296lcVEq5tRsn9iAr/wSPfrTN1aFwqKiMTzblcnl6Ep07NZKwju6GRTfBgovqThDJ4+GWryBtjiaIOui9m5RSTTK5XyzXj+/B81/vZULvaKaluuZ22MYYfv/uZgCubajBOmctLP87bH0PTPWZ84O6wLkPw9ArNTk0QEsSSqkmu3tmP1Ljw7jrzfUcctHFdS+t3MfHm3P59Yx+JHcJPn2mMdbtvBfMgnmTYcs7dSeItLlwRwYMu0oTRCM0SSilmizAz5d/XjWcsopq7nx9HVXVpvEXOdHmnEIe+mArU/rFcOOEnt/PqCiDda9YYz0smGUlirrEDYYbPocL/wFBUe0Ss6fT6ialVLP0jg3hwYsG8utFG3hm2W5um9y7XbZ7vLySn7yylshgfx6/fJg1fnVBljXO9JoXzhznwVHsQJjwcxg4G3z1tNccerSUUs32g/RElu7M4/FPdzCmZxdGdG/mdQot8Nt3NpF59Div3DiaqMMr4dt5sO3DuquTaiSPh/E/t27Ip9VKLaJJQinVbCLCI5cMZt3+An722lo+/OlEwgLbbtjON7/L5q21B7j3rFjGfHwB5G1tKDroNxMm3AlJ2hu/tbRNQinVIuGd/fnnVcPIKSjj3rc2YkzbtE/sOlzMb9/ZxNieXbjx3LT6x5AOCLdupfGT7+CqVzVBOImWJJRSLZaWHMVd0/vx6Efb6B8Xyk+m9nHKeo0x5BaVsflAEX/5ZDtBnXz5+5XD8PX1gdG3wLu3f79wbCqMuhmGXA6dgutfqWoRTRJKqVa5ZVJPtucW8/hnO+gdG8LMwfHNXkd1teGzrYdYs/8YW3KK2JxTRP7xkwB08vPh2evSiQuz7/I66FL4/AFIHmclh+Tx2t7QhjRJKKVaRUT4v9mDyTx6nF8sXE9SVBCDujV93OfNOYX85u1NrMsqwN9X6BsXyrQBsQxMCGdgQhgD4sMIDnA4Vfl3hp+t11JDO5G2qkdsjvT0dJORkeHqMJRSrXC4uIyLn/gaA7x7+3hiwxoe36G4rIK/fraDF77JJDKoE/edN4ALhya41a3I3Z2IfGeMSW/Lbei7oZRyitjQQJ6dk05BaQU3v/gdZRVVdS5njOGDDTlM++tS5n+TyVWjuvPlLydzaVqiJgg3pNVNSimnGZgQzt+uGMqtL63h7kUb+PsVwygur2TnoRJ2Hipm5+ES1uw/xtr9BQxMCOOZa9MZlhTh6rBVAzRJKKWcasageH51bl8e+3QHy3ce4ajdAA0Q6O9D79gQHrgwlWvGJOPnqyUHd6dJQinldLdPsW7VsTvvOH3iQugbG0qfuBASI4Pw9dGeSJ5Ek4RSyulEhDvOds41E8q1tKynlFKqXpoklFJK1UuThFJKqXppklBKKVUvTRJKKaXqpUlCKaVUvTRJKKWUqpcmCaWUUvXSJKGUUqpemiSUUkrVS5OEUkqpemmSUEopVS9NEkopperlFsOXikgxsN0JqwoHCt1gHc5cTzRwxE1icaf1OOO4OCsWdzouoJ+Z+njjZ6aPMabpA4q3hDHG5Q8gw0nrmecO63Dyelp9bNxwn5zxPulnpg2PjRvuk35m2nA9DT28rbrpfTdZhzPX4wzutk/edmz0uHjOepyhQ31m3KW6KcMYk+7qONyRHpu66XGpnx6buulxaRl3KUnMc3UAbkyPTd30uNRPj03d9Li0gFuUJJRSSrkndylJKKWUckNtkiREJElEFovIFhHZLCI/s6dHichnIrLT/htpT+8vIitEpFxEfuWwnn4iss7hUSQiP2+LmNuLs46NPe9Oex2bRORVEQl0xT45g5OPy8/sY7LZ0z8v0KJjc7WIbBCRjSLyjYgMdVjXDBHZLiK7ROQeV+2TMzj5uDwvIodFZJOr9sdttUWXKSAeGGE/DwV2AKnAn4F77On3AH+yn8cCI4GHgV/Vs05fIBdIbusuX235cNaxAboBe4HO9v8Lgbmu3j83OC6DgE1AEOAHfA70dvX+tfOxGQdE2s9nAqvs577AbqAn0AlYD6S6ev9cfVzs/ycBI4BNrt4vd3u0SUnCGHPQGLPGfl4MbMU6qc0CXrAXewG42F7msDFmNVDRwGqnAruNMfvaIub24uRj4wd0FhE/rJNiTttG33aceFwGYH35S40xlcBSYHbb70HbacGx+cYYc8yevhJItJ+PAnYZY/YYY04Cr9nr8EhOPC4YY5YB+e0TuWdp8zYJEUkBhgOrgDhjzEF7Vi4Q14xVXQm86tzoXKs1x8YYcwB4DNgPHAQKjTGftl207aeVn5lNwEQR6SIiQcB5QFJbxdreWnBsbgA+sp93A7Ic5mXb0zxeK4+LakCbJgkRCQEWAT83xhQ5zjNWGa9JXatEpBNwEfCG04N0kdYeG7uedRbQA0gAgkXkmjYKt9209rgYY7YCfwI+BT4G1gFVbRJsO2vusRGRKVgnw7vbLUgX0OPSttosSYiIP9Yb97Ix5i178iERibfnxwOHm7i6mcAaY8wh50fa/px0bKYBe40xecaYCuAtrDpXj+Wsz4wx5jljTJoxZhJwDKuu2qM199iIyBDgP8AsY8xRe/IBTi9VJdrTPJaTjotqQFv1bhLgOWCrMeavDrPeA+bYz+cA7zZxlVfhJVVNTjw2+4ExIhJkr3MqVp2sR3LmZ0ZEYu2/3bHaI15xbrTtq7nHxt7vt4BrjTGOCXI10EdEetil8yvtdXgkJx4X1ZC2aA0HJmAV8TZgFffXYdUNdwG+AHZi9TqJspfvilU/WgQU2M/D7HnBwFEgvD1b9Nvq4eRj8yCwDase/kUgwNX75ybH5StgC1bvnamu3jcXHJv/YJWgapbNcFjXeVglq93Ab1y9b250XF7FatursD9LN7h6/9zloVdcK6WUqpdeca2UUqpemiSUUkrVS5OEUkqpemmSUEopVS9NEkoppeqlSUIppVS9NEkopZSqlyYJpZRS9fp/6TRyda/rkFcAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "dfgas['gasrig_count'].loc['2017':].plot(legend=True)\n",
    "dfgas_forecast['gasrig_forecast'].plot(legend=True,linewidth=4,title='Gas Rig Count Prediction: FORECAST vs TEST Set - VAR Model');"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "0a71e73a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Gas Rig VAR(4) RMSE: 11.890\n",
      "Gas Rig VAR(4) MSE: 141.381\n",
      "Gas Rig VAR(4) MAE: 11.385\n",
      "\n",
      "Gas Rig Mean Value: 94.254\n",
      "Gas Rig std Value: 7.984\n",
      "Gas Rig Percent Change: 12.615%\n"
     ]
    }
   ],
   "source": [
    "RMSE2 = rmse(dfgas['gasrig_count'].loc[test.index.min():],dfgas_forecast['gasrig_forecast'])\n",
    "MSE2 = mse(dfgas['gasrig_count'].loc[test.index.min():],dfgas_forecast['gasrig_forecast'])\n",
    "MAE2 = meanabs(dfgas['gasrig_count'].loc[test.index.min():],dfgas_forecast['gasrig_forecast'])\n",
    "\n",
    "print(f\"\"\"Gas Rig VAR(4) RMSE: {RMSE2:.3f}\n",
    "Gas Rig VAR(4) MSE: {MSE2:.3f}\n",
    "Gas Rig VAR(4) MAE: {MAE2:.3f}\\n\n",
    "Gas Rig Mean Value: {dfgas['gasrig_count'][-nobs:].mean():.3f}\n",
    "Gas Rig std Value: {dfgas['gasrig_count'][-nobs:].std():.3f}\n",
    "Gas Rig Percent Change: {(RMSE2/dfgas['gasrig_count'][-nobs:].mean())*100:.3f}%\"\"\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "b66a8d5e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "TRAIN: 178 TEST: 12\n",
      "TRAIN: 190 TEST: 12\n",
      "TRAIN: 202 TEST: 12\n",
      "TRAIN: 214 TEST: 12\n",
      "TRAIN: 226 TEST: 12\n",
      "TRAIN: 238 TEST: 12\n",
      "TRAIN: 250 TEST: 12\n",
      "TRAIN: 262 TEST: 12\n",
      "TRAIN: 274 TEST: 12\n",
      "TRAIN: 286 TEST: 12\n",
      "\n",
      "10-fold Cross Validation Gas Rig\n",
      "RMSE: 190.620,\n",
      "MSE: 66785.434\n",
      "MAE: 187.414\n",
      "\n",
      "Gas Rig Mean Value: 94.254\n",
      "Gas Rig std Value: 7.984\n",
      "Gas Rig Percent Change: 202.240%\n"
     ]
    }
   ],
   "source": [
    "tscv = TimeSeriesSplit(n_splits=10, test_size=12)\n",
    "RMSE_gas = []\n",
    "MSE_gas = []\n",
    "MAE_gas = []\n",
    "for train_index, test_index in tscv.split(dfoil_transformed):\n",
    "    print(\"TRAIN:\", len(train_index), \"TEST:\", len(test_index))\n",
    "    train,test = dfgas_transformed.loc[dfgas_transformed.index[train_index],:], dfgas_transformed.loc[dfgas_transformed.index[test_index],:]\n",
    "    gas_model = VAR(train)\n",
    "    gas_results = gas_model.fit(4,trend='ctt')\n",
    "    z = gas_results.forecast(y=train.values[-lag_order:], steps=nobs)\n",
    "    idx = pd.date_range(test.index.min(), periods=nobs, freq='M')\n",
    "    dfgas_forecast = pd.DataFrame(z, index=idx, columns=['gasrig_count2d', 'gas_price2d', 'gaswell_count2d', 'gasimports_amount2d',\n",
    "       'gasconsumption_amount2d'])\n",
    "    dfgas_forecast['gasrig_count1d'] = (dfgas['gasrig_count'].iloc[-nobs-1]-dfgas['gasrig_count'].iloc[-nobs-2])+dfgas_forecast['gasrig_count2d'].cumsum()\n",
    "    dfgas_forecast['gasrig_forecast'] = dfgas['gasrig_count'].iloc[-nobs-1] + dfgas_forecast['gasrig_count1d'].cumsum()\n",
    "    RMSE_gas.append('{:.3f}'.format(rmse(dfgas.loc[dfgas.index[test_index],'gasrig_count'],dfgas_forecast['gasrig_forecast'])))\n",
    "    MSE_gas.append('{:.3f}'.format(mse(dfgas.loc[dfgas.index[test_index],'gasrig_count'],dfgas_forecast['gasrig_forecast'])))\n",
    "    MAE_gas.append('{:.3f}'.format(meanabs(dfgas.loc[dfgas.index[test_index],'gasrig_count'],dfgas_forecast['gasrig_forecast'])))\n",
    "\n",
    "RMSE_gas = np.array(RMSE_gas).astype('float')\n",
    "MSE_gas = np.array(MSE_gas).astype('float')\n",
    "MAE_gas = np.array(MAE_gas).astype('float')\n",
    "print(f\"\"\"\\n10-fold Cross Validation Gas Rig\n",
    "RMSE: {RMSE_gas.mean():.3f},\n",
    "MSE: {MSE_gas.mean():.3f}\n",
    "MAE: {MAE_gas.mean():.3f}\\n\n",
    "Gas Rig Mean Value: {dfgas['gasrig_count'][-nobs:].mean():.3f}\n",
    "Gas Rig std Value: {dfgas['gasrig_count'][-nobs:].std():.3f}\n",
    "Gas Rig Percent Change: {RMSE_gas.mean()/dfgas['gasrig_count'][-nobs:].mean()*100:.3f}%\"\"\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ea0ea120",
   "metadata": {},
   "source": [
    "RE-Fit model with 100% of data for TRUE Forecasting"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "3dabb447",
   "metadata": {},
   "outputs": [],
   "source": [
    "gas_model = VAR(dfgas_transformed)\n",
    "gas_results = gas_model.fit(4,trend='ctt')\n",
    "z = gas_results.forecast(y=dfgas_transformed.values, steps=nobs)\n",
    "idx = pd.date_range(dfgas_transformed.index.max(), periods=nobs, freq='M')\n",
    "dfgas_forecast = pd.DataFrame(z, index=idx, columns=['gasrig_count2d', 'gas_price2d', 'gaswell_count2d', 'gasimports_amount2d',\n",
    "       'gasconsumption_amount2d'])\n",
    "dfgas_forecast['gasrig_count1d'] = (dfgas['gasrig_count'].iloc[-1]-dfgas['gasrig_count'].iloc[-2])+dfgas_forecast['gasrig_count2d'].cumsum()\n",
    "dfgas_forecast['gasrig_forecast'] = dfgas['gasrig_count'].iloc[-1] + dfgas_forecast['gasrig_count1d'].cumsum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "336bdb82",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "dfgas['gasrig_count'].loc['2017':].plot(legend=True)\n",
    "dfgas_forecast['gasrig_forecast'].plot(legend=True,linewidth=4,title='Gas Rig Count Prediction: REAL FORECAST - VAR Model');"
   ]
  }
 ],
 "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": "3.9.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}