Chapter 2 - Supervised Learning - Regression.ipynb

{
 "metadata": {
  "name": "",
  "signature": "sha256:15d40bfd4f11687d7efb62373653c0fcdf77c904fcecd2424694aaeeaff94849"
 },
 "nbformat": 3,
 "nbformat_minor": 0,
 "worksheets": [
  {
   "cells": [
    {
     "cell_type": "heading",
     "level": 1,
     "metadata": {},
     "source": [
      "Learning Scikit-learn: Machine Learning in Python"
     ]
    },
    {
     "cell_type": "heading",
     "level": 2,
     "metadata": {},
     "source": [
      "IPython Notebook for Chapter 2: Supervised Learning - Estimating Boston house pricing using Linear Regression"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "_In every example we have seen so far, we have faced what in Chapter 1, Machine Learning \u2013 A Gentle Introduction, we called classification problems: the output we aimed at predicting belonged to a discrete set. But often, we would want to predict\n",
      "a value extracted from the real line. In this notebook, we want to approximate a real function (i.e. solve a _regression problem_), using Linear Regression, a very simple regression model, from Scikit-learn._"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "Start by importing numpy, scikit-learn, and pyplot, the Python libraries we will be using in this chapter. Show the versions we will be using (in case you have problems running the notebooks)."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "%pylab inline\n",
      "import IPython\n",
      "import sklearn as sk\n",
      "import numpy as np\n",
      "import matplotlib\n",
      "import matplotlib.pyplot as plt\n",
      "\n",
      "print 'IPython version:', IPython.__version__\n",
      "print 'numpy version:', np.__version__\n",
      "print 'scikit-learn version:', sk.__version__\n",
      "print 'matplotlib version:', matplotlib.__version__"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Populating the interactive namespace from numpy and matplotlib\n",
        "IPython version: 2.1.0\n",
        "numpy version: 1.8.2\n",
        "scikit-learn version: 0.15.1\n",
        "matplotlib version: 1.3.1\n"
       ]
      }
     ],
     "prompt_number": 1
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "Import the Boston House Pricing Dataset (http://www.cs.toronto.edu/~delve/data/boston/bostonDetail.html), and show their features."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from sklearn.datasets import load_boston\n",
      "boston = load_boston()\n",
      "print boston.data.shape\n",
      "print boston.feature_names\n",
      "print np.max(boston.target), np.min(boston.target), np.mean(boston.target)\n",
      "print boston.DESCR"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "(506, 13)\n",
        "['CRIM' 'ZN' 'INDUS' 'CHAS' 'NOX' 'RM' 'AGE' 'DIS' 'RAD' 'TAX' 'PTRATIO'\n",
        " 'B' 'LSTAT']\n",
        "50.0 5.0 22.5328063241\n",
        "Boston House Prices dataset\n",
        "\n",
        "Notes\n",
        "------\n",
        "Data Set Characteristics:  \n",
        "\n",
        "    :Number of Instances: 506 \n",
        "\n",
        "    :Number of Attributes: 13 numeric/categorical predictive\n",
        "    \n",
        "    :Median Value (attribute 14) is usually the target\n",
        "\n",
        "    :Attribute Information (in order):\n",
        "        - CRIM     per capita crime rate by town\n",
        "        - ZN       proportion of residential land zoned for lots over 25,000 sq.ft.\n",
        "        - INDUS    proportion of non-retail business acres per town\n",
        "        - CHAS     Charles River dummy variable (= 1 if tract bounds river; 0 otherwise)\n",
        "        - NOX      nitric oxides concentration (parts per 10 million)\n",
        "        - RM       average number of rooms per dwelling\n",
        "        - AGE      proportion of owner-occupied units built prior to 1940\n",
        "        - DIS      weighted distances to five Boston employment centres\n",
        "        - RAD      index of accessibility to radial highways\n",
        "        - TAX      full-value property-tax rate per $10,000\n",
        "        - PTRATIO  pupil-teacher ratio by town\n",
        "        - B        1000(Bk - 0.63)^2 where Bk is the proportion of blacks by town\n",
        "        - LSTAT    % lower status of the population\n",
        "        - MEDV     Median value of owner-occupied homes in $1000's\n",
        "\n",
        "    :Missing Attribute Values: None\n",
        "\n",
        "    :Creator: Harrison, D. and Rubinfeld, D.L.\n",
        "\n",
        "This is a copy of UCI ML housing dataset.\n",
        "http://archive.ics.uci.edu/ml/datasets/Housing\n",
        "\n",
        "\n",
        "This dataset was taken from the StatLib library which is maintained at Carnegie Mellon University.\n",
        "\n",
        "The Boston house-price data of Harrison, D. and Rubinfeld, D.L. 'Hedonic\n",
        "prices and the demand for clean air', J. Environ. Economics & Management,\n",
        "vol.5, 81-102, 1978.   Used in Belsley, Kuh & Welsch, 'Regression diagnostics\n",
        "...', Wiley, 1980.   N.B. Various transformations are used in the table on\n",
        "pages 244-261 of the latter.\n",
        "\n",
        "The Boston house-price data has been used in many machine learning papers that address regression\n",
        "problems.   \n",
        "     \n",
        "**References**\n",
        "\n",
        "   - Belsley, Kuh & Welsch, 'Regression diagnostics: Identifying Influential Data and Sources of Collinearity', Wiley, 1980. 244-261.\n",
        "   - Quinlan,R. (1993). Combining Instance-Based and Model-Based Learning. In Proceedings on the Tenth International Conference of Machine Learning, 236-243, University of Massachusetts, Amherst. Morgan Kaufmann.\n",
        "   - many more! (see http://archive.ics.uci.edu/ml/datasets/Housing)\n",
        "\n"
       ]
      }
     ],
     "prompt_number": 2
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "Show an example instance, and some common statistics:"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "print boston.data[0]\n",
      "print np.max(boston.data), np.min(boston.data), np.mean(boston.data)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "[  6.32000000e-03   1.80000000e+01   2.31000000e+00   0.00000000e+00\n",
        "   5.38000000e-01   6.57500000e+00   6.52000000e+01   4.09000000e+00\n",
        "   1.00000000e+00   2.96000000e+02   1.53000000e+01   3.96900000e+02\n",
        "   4.98000000e+00]\n",
        "711.0 0.0 70.0724468258\n"
       ]
      }
     ],
     "prompt_number": 3
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "Build, as usual, training and testing sets:"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from sklearn.cross_validation import train_test_split\n",
      "X_train, X_test, y_train, y_test = train_test_split(boston.data, boston.target, test_size=0.25, random_state=33)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 4
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "Before learning, let's try to see which features are more relevant for our learning task, i.e. which of them are better prize predictors. We will use the SelectKBest method from the feature_selection package, and plot the results."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from sklearn.feature_selection import *\n",
      "fs=SelectKBest(score_func=f_regression,k=5)\n",
      "X_new=fs.fit_transform(X_train,y_train)\n",
      "print zip(fs.get_support(),boston.feature_names)\n",
      "\n",
      "x_min, x_max = X_new[:,0].min() - .5, X_new[:, 0].max() + .5\n",
      "y_min, y_max = y_train.min() - .5, y_train.max() + .5\n",
      "#fig=plt.figure()\n",
      "#fig.subplots_adjust(left=0, right=1, bottom=0, top=1, hspace=0.05, wspace=0.05)\n",
      "\n",
      "# Two subplots, unpack the axes array immediately\n",
      "fig, axes = plt.subplots(1,5)\n",
      "fig.set_size_inches(12,12)\n",
      "\n",
      "for i in range(5):\n",
      "    axes[i].set_aspect('equal')\n",
      "    axes[i].set_title('Feature ' + str(i))\n",
      "    axes[i].set_xlabel('Feature')\n",
      "    axes[i].set_ylabel('Median house value')\n",
      "    axes[i].set_xlim(x_min, x_max)\n",
      "    axes[i].set_ylim(y_min, y_max)\n",
      "    sca(axes[i])\n",
      "    plt.scatter(X_new[:,i],y_train)\n"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "[(False, 'CRIM'), (False, 'ZN'), (True, 'INDUS'), (False, 'CHAS'), (False, 'NOX'), (True, 'RM'), (False, 'AGE'), (False, 'DIS'), (False, 'RAD'), (True, 'TAX'), (True, 'PTRATIO'), (False, 'B'), (True, 'LSTAT')]\n"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAsoAAAD2CAYAAADPsaG/AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzsnXmczdX/x593m7vNwswY69j3fV8iW4pSUUqIhGhBSpIk\nkYo2bbSgskR9fyUqSpslhRBSdtmzZJ99ufe+f398rnGZxYyZuzrPx2Me7v18zuec9zn35X3O53ze\nn3NAoVAoFAqFQqFQKBQKhUKhUCgUCoVCoVAoFAqFQqFQKBQKhUKhUCgUCoVCoVAoFAqFQqFQKBQK\nhUKhUCgUCoVCoVAoFAqFQqFQKBQKhUIRLBwAUoBE918CUKoI8uxYyDwKyuPAMeA88CEQ5uPyrwUO\nEPxaqQt8D5wEXD4s91rjAMGvlf7ARjSfchh4GTD4sPxrgQMEv056ATvRdHIK+BIo48PyrxUOEPxa\n8eRntD5I76fycyXgDAoABLgViHD/RQLHiyBPXSGuL2hn1Bl4Ck3wFYDKwMRClK/ImVDQSgbwGTCo\nEGUqrkwoaMUKjABigBbADcCoQpSvyE4o6OQ3oC0Qhdb/pABTC1G+ImdCQSsXuBcwustXBAH7yfmO\nKgptZvYocASYxMUbjSrAcrS755PAJ+70APMAJxfv/EYB7dFmZDw54FHuBOAL97XngYFXKP9yFgAv\neHzvgDa7rChaQkErF6iKmlH2JqGklQs8Dnydz7SK/BFqOgkH5gBv5COtomCEilaigF1oN99qRjmI\nyOmOajba7FsVoBFwE/CAx/kXgdJALSAeTUAA/YBDXLzzey2XMi+/k7od+BxNRAvyUb4ntYE/Pb5v\nBUoCxXNJr7h6gl0rCt8RalppB/ydz7SK/BMKOmkDnEMLByiP9oRTUfSEglZeAt4FTuSRRhFgHEC7\nmzrr/vsSbZCZBlg80vVGuzPLie7AJo/vl9/5tSf7XZpnmgnASo9zBS1/L5o4L2BCu1Mrn0t6xdVx\ngODXygXUjLJ3OUDoaAW0maNDQHQ+0iryzwFCSydlgB+At/KRVlEwDhD8WmnqLl8PVCRAZ5SN/jYg\nABGgG5f+sM3RBpue4Qt6tI4CNHG8hXYXHeE+d6aQdhzx+FzhCuVfThJavNIFLjxaSSykTYpLCQWt\nKHxDKGmlO9os0A1FYI/iUkJJJ6A9fn8WWIYW364oOoJdK3q0meTHuHSSpjAx0l5BDZTzx2EgHe0l\nlpxm3V5Ci+2pi/a4qTvwjsf5yx9VJAM2j+8GoMRlaTyvuVL5l7MNaIgWOwTQAO2xxtl8XKsoHMGm\nFYX/CEatdAFmALeg+RmF9wlGnXhiQot7VXifYNJKJNAE+J9H3qANvO9Ceyk0IAi4Ke4A5Rja46Op\nXLwLq4L2Zi9oLywko8VjlQWevOz6E+70F9iN9mjiFjQnMg4wF6L8y5mLtopBLbS45GeBj/OuoqKI\nCDat4M7/wvKB5ivkryg6gk0rHYH5wJ1oy8QpfEOw6aQPWuwraDOMLwIL88hfUXQEk1bOocVKN3D/\n3eI+3hhYn2ctfYwaKOef+9AGE9vRHlV8zsU1Cyei/bjngW/QnILnXdZkNIGdBUa60z0CzEK7e0ri\n0jggIfudXV7lX873wCvACrQ4pn+A5/JfVUUhCSatVESb7fnbnU8qsCP/VVUUkmDSyji0zu87Lq7d\nurQAdVVcPcGkk9rAGne+K4G1wOj8V1VRSIJJK/95/J1y53UCyMx/dRUKhUKhUCgUCoVCoVAoFAqF\nQqEIFHzxduEBtHgYJ9p0enO0JYX+hxa/dADoiRavorh2OYDSiSJ/HEBpRXFlDqB0osgfB1BaUfiR\n/WRfa/MVLsYsPQVM8alFikBE6USRX5RWFPlB6USRX5RWFH5lP9pSIZ7sRFvPD7Qg750+tUgRiCid\nKPKL0ooiPyidKPKL0orCr+wDNqMtJzTYfcxzPV8dOazvq9OFXXibUv0F/99KrsxV6aRBgwb+rpv6\nU1pRfyGkk3bt2vm7buovSLSifEpI/a3Ej5R2/1sC2AJcT3bB5bQzjIAIiBiND8vkyZPlAu3a3Sbw\nSdZ5eEegkcTFxQuM8Tj+s5hMJURE5LnnnhMRkbfeekvM5v4eaXZKVFQpuRou5FlU5Dc/iBTY4FGH\nh6V48WiBYgIrPY4/KXZ7pF9svNRexJs6KQoiIkoI7HG323MCd7l1JQKxAgMEJrq/nxewCSS5v7sE\nmggsc3+fIRER5QvVZlfCX9rzdp7BoJXC1M/XeYaqjd7WSSDUsbD5HTx4UCDsMj9VT3r16iVmcyeP\nfuKUgFkgUvr3758tn2PHjonZXFwg1SOfBgIfSYkSFbOl1+liBGa6/agIzBKdLroIah2YWvHkySef\nFhjt0bYrBKwCgz2ObRaIcn8+7m772wXmeLTZu3L77X0uyXvdunUSHl5bwOlOkyYWSwmZPn26hId3\n8cj/tBiNZsnMzLzqNsuLa9Gn+GJnvgtbGZ4EFqEFyZ9Ae5RxHE2g/+V86QQAnM7N7N6dnnU0M9MB\nWD3SWQE9TqcA9kuOa/W/iMPhQOTSa51OR8Fq5HeES+tvc9edbMcvr78vWLlyJStXrizoZVetkwkT\nJmR9bt++Pe3bty9o2W4NeLadHbigCwfahILV47sebf11PM5dSG/D5fJ9uwcjwagVhe/xtU5WrlyZ\npZVg1UlmZiaan7qwl5DmpzIyMhCxeKS0oPUpBtLT07kch8OBTmfk4nBB577GgMORve8U8fSVoPXD\nvtuV2J8+Zd++fWj7fF3Aita2thyOeX7OPqbRxjkX0X4HCxfXYDCi15vIyMi47FoLIoLLpTZmzYuC\n6MTbA2Ub2raEiWgjj5vQFrz+GugPvOz+d3HOl7cB9mOz7WX06I+zjg4f3p8NGx4lM9MIZABjMZvT\nGDFiFOPHTwHKo4UWDaVz5+svyfGOO+7guedakJFRD6iBzfYcAwbcX3Q19gERETYSE+8B3gAOAu8x\nYsQT/O9/i9m9+15gGtr/7Vdp2bKlz+27vGOZOHHilS4plE48HdXVMmDAAD7+uDcpKc8B69HpfsFg\nCMPhWIleXwGXax3wFVAG7Z0PC9ALGA78hPbU7hyajx1B//69C23TtUAwakXhe3ytk/bt2we9VipV\nqkTx4nGcPdsTeBT4Gb1+BxMmzGLFipvJyJgMtER7Z60msIPJkydny6ds2bI0atSAP/64n4yMB4Bv\ngRNYre8wePD92dJfd1111qwZAdyA5g8fpVWr6t6qZjb86VM2bdrEN9/cQEZGRbQx9uNoA+EZaPuw\nVEbbx6MM2n5gE9zF13WnbQMsxGZ7hmHDZl5STpMmTYiNzSAtbTSZmV0xm+dSu3Y1+vTpw6RJr6HX\nv4zL1Ryr9XW6dr2HsLAwFLlzFTrxGpXQHmNsQdv562n38Wi00cVutO0Oi+VwrTRs2E46dLhdNmzY\nkG2afP78BVK+fD2x2cpJ7dpNZdWqVSIi8uijj4rBUEJ0ulhp2LC5OBwOERFZsWJF1rVbtmyRTp26\nS4MGbWXChJey0hQUzzyLgvzml5qaKrGx5QSiBYrLgAEDRETE4XBI/fpNRaeLFYOhhIwYMcJvNnrC\nlR99FUonRYHD4ZDnnntBGjRoK02atJEVK1ZIz573S716beT++x+Wp54aJ5Ur15OoqIoSH19XSpas\n7H58VtwdCmMSnS5GLJYy8vDDQwvdZlciEH5Xb+QZDFopTP18nWeo2uhtnQRCHYsiv4MHD0p0dHnR\n6WLFbI6Tb775RkRE9uzZI61b3+TuQyJFry8mEydOzDWfhIQEGTx4uFSt2kRKlKgideq0lJdeelWc\nTmeO6Zs2ber2i9HSuHHTAtcvNwJRK5fzyy+/SLVqTcRmKycVK9aR7777Tu6+u7fo9TECMWIyRYnJ\nFCsWSxlp2rSN1K7dWKzWslK8eDkpU6a6NG9+o7z22mtyww3dpVGjDjJ58mtZ7Xz8+PGsfmngwKFy\n/vx5ERHZu3ev3HJLT6lf/3oZOfJpSUtLK1Sb5cW16FN89zyk4Lhtz52NGzeycOFiwsNtDBo0kKNH\nj3L99Z1JSXkNiMNme5Lnnx/ME0+M8I3FihzR6XTgPa1dUSfewOVyMX/+fKZOncb27efIyHgbnW43\n4eEv8PffGyhfvrzPbQoFQlEriqJH6eTKiAg33dSdX381k5b2EAbDSuLiFrBr12YiIiIQET799FM2\nb95KjRpVGTBgAAaDwd9mFzmBqJXFixfz22/rqFgxngceeACz2Zx1bvfu3TRu3Jrk5BeASthsz/Lo\no12YPNl/M57XAl7WidfIc/S/bNkysVpLCIwTk2mIxMbGy6BBDwo87xHUvkYqVWqQ7dqdO3dK9+73\nSuvWt8jUqW+Ly+Uq8N2HIv+Qv5cpvKITb9G372Cx25sJvOB+ka+SwBwxmwfJW2+95RebQoFQ1Iqi\n6FE6uTKnT5+WsLAIgfSsPjEior0sXbpUREQeeGCY2O2NBV4Qm62t3HLLXbn2hSkpKTJy5NPSqlUX\nGTDgETl16pQvq1IoAk0rzzwzUazW6gKTRK/vKBUr1pGMjIys8y+88KIYDI95jGN2SvHi5YqySRQ5\nkJdOfPEyn1cYOXICqamzgNvJzIRz54bz11+b0OliuHiDl5btDvnw4cM0b96OxMSRiNRg8+YXOHHi\nJFOmPO/rKiiClIMHD/LFF4tIS9sPhANPoMWePYvTWRqjsbF/DVQoFNc8BoMBERfaRnNhaOMArU88\nduwY8+Z9Qnr6QSCSlJRRrFpVi82bN9O48aX+S0S49daerFljJi1tKBs3LmP16hv46691WCyW7AUr\nciUtLY0pUybjdB4ASuJyjeXAgToMGjSEuXO197CMRgN6fRpOZ9ZVGI1BO1QLCfT+NuBqSUxMBOKz\nvjsc5alSpRo22wx0uinAx9hsAxg37rFLrvviiy9IS7sNkdFAN1JS/se7777vU9sVwU1iYiJGYzTa\nIBm0F/tKA2NxufZw1113+c84hUKhAKKiorjzzrux2W4H5hMW9iClSqXRrl07kpKSMBqjgAh3ajNG\nYymSkpKy5XP06FHWrFlLWtqnwK1kZr7DiRNG1q1b58PahAapqamI6NFWoQNtCFaVRYuWZKW59957\nsdkWo9dPAOZis93DU089lj0zhc8I2oFyz57dsNlGom2W8wtW65s88MB9/P77Svr0+Ydu3X7mk0/e\npH//fpdcp9Pp0Ok8Z9iFIAxLUfiR6tWrEx2tR6ebhLah0xtoS25WJD6+LHFxcf41UKFQKIBPPpnJ\n+PG3cMst3/Dww8VYv34FFouFSpUqERcXgcEwEdiPTjcNk+lfGjZsmC0Pd+wmlz6ZFo/jivxSrFgx\nSpUqBwxD6zvmAb8TFnZxZr5cuXJs2vQb9913gltvXcb77z+r3rNS5Eqe8SSZmZkyYsRoKVGikpQv\nX0cWLPg0X3Eohw8flqioUqLXvyTwhdhsjWTs2OcKH+CiyBUCLEasKDhw4IA0anS9QLhAHYE3xWar\nIu+/P8Mv9oQKoagVRdGjdFJ4Dh8+LO3adZXixctJ06YdZOfOnTmmc7lc0qnT7WK1dhdYJGFhD0m1\nag0vWVkhkAk0rWzcuFEMhuICJQRqi8VSVSZNmuKFmisKQl46CeRbQrftRc+ePXt45pkX+e+/M/To\n0YVhwx5Wd8deJBDfOi4q/vjjD55/fipJSSkMHNiTe+9V6ycXhlDWiqLoUDrxLWlpaUyY8BK//baJ\n2rWrMGXKBIoXL+5vs/JFIGplx44djBs3mTNnztOr120MGTJIjUH8TF46CeRfxqvO6siRI5w9e5Zq\n1aqpFxK8TCA6qsKSmZnJ7t27sdvtVKxY0eflhyqhqBVF0aN04nsSEhLYv38/5cqVIyYmxt/m5JtA\n1IqIsG/fPjIyMqhevXpILssXbOSlk6CNUb5aRIThw5+kWrUGtG7dk0qV6rB7925/m6UIIo4cOUL1\n6o1o2fIOatVqTp8+g9R2oQqFImT58ccfKVu2Cm3b3ku5clX58MPZ/jYpaMnIyODmm3tQr971NGvW\nlQYNruP06dP+NkuRB9fcQPmbb77h44+/JS1tL4mJOzhx4nHuuut+f5ulCCL69XuYw4fvJilpF2lp\n+/n66+3MnTvX32YpFApFkZOamsqdd/YhKekLEhL+Ji3td4YPf5L9+/f727SgZOrUt/jll1RSUw+Q\nnPwPu3c3Z+jQJ/1tliIPQmagvG3bNhYuXMhff/11xXRpaV0BLb5KpA+7d//tAwsVocK2bdtwOtsC\ni4A/SU7uxubNSkMKhSL4WL9+PQsXLmTfvn05nj969CgidqCd+0h1wsIasGvXLp/ZGApcaOeVK9eS\nmno32trWOjIze6v+I8AJiYHyW29Np1mzGxg4cB4tW3Zm8uTXck1bvXp1LJYfgQvrRS6mUqUaPrFT\nERrExhYH7gDmAPdhMHxAnTrV/WyVQqFQFIyHHnqMjh3vYeDAedSt24KFC7/MlqZ06dKIJAAb3EcO\nkpGxlapVq/rU1mDGs52XL1+ByfQB4AAEk2kxdeqoMYji6sjXkh4nTpwQi6WYwAH3do//isUSIwcP\nHswxvcvlkn79hojNVkaioppJdHQ52bp169WtJ6LIFwTY8jyFweVyid0eI7DGrbckMRgqyvfff+9T\nO0KVUNKKwnsonRSeNWvWiM1WSeC825f9IVZrlDgcjmxpFy/+Smy2GImKailWa4y8+eY0P1h8dfhb\nK7/99lu2dtbrLWKzVZaIiHpSqVJdOX78uA9aQpEXeekk6PdFPHr0KEZjHPAIsA2ogclUliNHjlC+\nfPls6XU6HXPnfsDYsTs5c+YM9erVIyIiIls6hSIn0tLSSEtLBGxAW+AQImHs2LGDm266yc/WKRQK\nRf44ePAgUBnoChwGWuJwuDh37ly2VS26dbud/fu3s3v3bipUqEB8fHwOOSpy4tChQxgMTYBI9xEr\nImaMxjRq1KjA559/TMmSJf1pouIKBH3oRXx8PMnJx4CWwM9AJxIT913xP3LNmjW57rrr1CBZUSCs\nVitly1YGOgL9gZ9xuW7k/ffnIWpJKYVCESTEx8eTkrIWGITWdxYDTERHR+eYPi4ujjZt2qhBcgFp\n1KgRDscq4G/gHNAOkRdISPiFzZub06VLD7VqUoAT9APlY8eOYbHEAc8CVYAnsdkqcuzYsVyvSU9P\nZ8GCBUyfPp1t27b5ylRFiDBmzKPodLXQOpiyQEv27t3O2rVr/WyZQqFQ5A+t72wH3I/Wd05HJI2E\nhIQc069evZp33nmHpUuXqkmBAlCjRg1mzHgTs/k6TKby6HQXtrC243BU5p9/9rJlyxZ/m6nIg6Af\nKIeHh7tfNEh2H0kDzhEeHp5j+rS0NFq16sSQITMZNWorzZt3YOnSpb4yVxEC1KhRA7s9CUgE2gOz\ncTh60KlTN3788Uf/GqdQKBT5IDw8HKPxJHBhNvMU4MpxA66XXnqVLl36MXr0Du65Zwz9+z+kBssF\noG/fPiQmnmbx4s+w2VKAHUBjYCWZmddz443dOHz4sJ+tVAQj+QrAdrlc0qfPILHbWwq8JDbb9dKt\nW29xuVw5pp81a5bYbDcJuNyB9culdOmqRRYQrsgOIfbiTWZmprRq1UlMptoCXTy09J1UqFDH5/aE\nEqGmFYV3UDopPBkZGdK0aTuxWrsKvCg2Wy0ZO3ZCtnRnz56VsLBwgX+zXmC22crLpk2b/GB1wQkk\nrTidTunQ4VbR6ysIvOhuTxGDYZz06zfESy2gyA956SToX+bT6XTMmzeDefPm8eef2yhXrjtdunQh\nMzOTsLCwbOlPnjxJRkY9Lu5UWJ9z50761GZFcKPT6Zg16y2GDh3GypUN8dTSmTNKSwqFIvAxmUys\nXr2MmTNncuDAEVq3foE777yT8+fPc/ToUcqXL4/dbufMmTOYTMXJyCjjvtKOyVSVkyeVrysImZmZ\n7Nu3j7lz36NNm1s5eLBh1jmnsz5Hj37qR+sUwUqB7whGjnxazOYoCQ+vLGXLVpO9e/dmS6MtiVNa\nYItAioSFPSQ33ti9KG5IFLlAAN3RF5aTJ09KnTrNxWaLl7Cw4mIwRLu1lCxhYQPlttt6+dSeUCOU\ntKLwHkon3mH+/E/Fai0mERHVJDw8Vn788UfJzMyU0qWriE43TSBd4GuJiIiTEydO+NvcfBEIWvnn\nn3+kXLnqEh5eWczmKGnVqqNYrW0E/hM4KjZbc3nzzXe83BKKvPCyTrxGgSq5ZMkSsdtrCpwSENHr\nX5cGDdrkmHb27LkSEVFCDAaTdOhwq5w+fboo2lmRCwSAoyoq7ryzn5hMw93hFikSFtZGwsLCxWAI\nkxtv7C7nzp3zqT2hRihpReE9lE7yT0ZGhhw8eFBSUlLyTHf48GGxWmMEtrpDAlZKeHisJCUlyc6d\nO6VGjSai1xukdOmq8uuvv/rI+sITCFpp3Lit6PWvutv1tNjttaRz59vEZLJJWJhdHn/8KXE6nV5u\nCUVe5KWToA+9uMDWrVtJS7sV0NZ/dLnuZ+vWZzh9+nS2NSH79+9H//79EBF0Ol0OuSkUObNx4xYy\nM2ejhVtYycjoQ+/eNZg/f6bSkkKhCCg2btxIly53kJrqwuVKYsaMd+nX794c0+7evZuwsDqkptZz\nH2kHFOPw4cPUrFmTnTs3qj7zKtmxYysu10L3t2iSk7ty3XURfPfdVwCqTQOcoF/14gJVq1bFaPwB\nSHEf+QaR4jzwwIhcr1HiVBQUERfwlfubA1jMyZNHlZYUCkVA4XQ66dLlDk6fvp+UlIdJSxvDgw8+\nxt69e3NMX6lSJTIytgOH3Ee24nSeokyZMllplJ+7OiyWSOBr97dUYBlbt25Dp9MVqE1/++03Xnzx\nRWbOnElaWpo3TFXkQMgMlO+66y5KlXICFYBGwONACr/+uvqSdImJiaxatYqNGzeqRb4VBcZqtQBv\nAHWBqsA6Dh06xsqVK3Ndf1ShUCh8zX///ce5c4nAZ2jLp35LRoadP//8M8f0lSpV4qWXnsNqbUJU\nVFus1o589NEHREZG5pjek0OHDvHzzz9z4MCBoqxCyFCmTGlgDNAaqAnE8e+/5wqUx0cfzeamm+5h\n/PgEHntsES1adCQ9Pd0L1iqCiQLHmHTseKPAHQIrBc4KfC42W5ms83v27JG4uIoSGdlK7Paq0qnT\n7ZKRkVF0QS6KHCEAYsSKirp1Gwv0ElgtsEFgsUAxiYq6TmJi4mXHjh0+tSfUCCWtKLyH0smVOXv2\nrIBJ4Kg7NjZDoILMnDkzz+v2798vy5cvlyNHjuSrnA8/nC1Wa4xERbUTqzVW3n8/7/x9TSBo5d57\nHxCDYZh7bLJFYIAYDFHyzjvv5bse4eGxHvHjLrHbO8iCBQuutlkUl5GXToJuRllE2L17N+vXryc5\nOfmSc5UrV0O7Y9Niq6AaUVEXt6ju338Yp04NJyFhDcnJ21mzJoWZM2f60nxFkHPddS3RZpOro4Ve\nxADFOX/+N86ceYq+fR/2q30KhUIB2g60JpMdKOU+YiIsrApxcXFZaU6ePMm6desu2cm2YsWKdOjQ\ngbJly16xjJMnTzJs2OOkpv7G+fMrSU1dx+OPP5XnzrjXIlOnvkjJkt8DA4F7gc04nSsZPXpcvjYa\ncblcpKaeR3uKCaDD5arKuXMFm5VWXB1BNVAWEfr1G0LDhu248cYHqVSpDjt37sw636PHbdhsbwN/\nAEewWEbTo8dtWef/+WcvLldX9zcTKSk3sX17zvFaCkVO9Op1FybTVKAGMBS4mQvOS6Qr+/YpPSkU\nCv8TFxdHlSpV0OvHA/8B/4fZ/DctWrQAYNGixVSsWJPOnYdTpUpdPvhgVoHLOHz4MCZTPJo/BKhC\nWFhlDh48WFTVCAlOnTpFSkoy2pDrJNAMaEBYWLV8havo9XrateuCyTQCOAH8iE63iPbt23vRakUw\nkG1q/LPPPhO7vYlAkoCITjdd6tdvfUmaGTNmSWxsBYmIiJOBA4dKenp61rkbb7xDjMYx7qW9EsRm\nay6zZ8/28oS+ggB49FVUaLtURQpsdj8C2ykQLXBYDIbx0q5dV5/aE2qEklYU3kPpJH/8+++/0rp1\nZ7HZoqVq1Yby+++/i4hIQkKC2GzFBTa6/dgesVpj5MCBAwXK/8yZM2K3xwisceezQWy2GDl58qQ3\nqnNVBIJW6tRpKTrd++42ShRoKPCK2GwxcuzYsXzlcebMGbn55rvEbo+RcuVqyrJlywrTLIrL8LJO\nvEa2ijz33ATR6Z7J2vYRjondHpPvhjh27JhUr95ILJaSotMVk2LFKsmzzz4vDoejKNtbcRkEgKMq\nKv7880+JiKjtocGTAiVFp4sRu720bN261af2hBqhpBWF91A6KRzbt2+XiIhqHn5MJCqqrfz88885\npj9//rwMHDhU6tS5Tu64o6/8+++/WeeWLFkqNlu0hIdXEZutuCxatNhX1cgX/tTK6tWrpXXrm0Wn\nixaY5J6kE4FRYjJZ5IsvFvqoFRRXwss68RrZKvJ///d/Yrc3FDjv3lTkDWncuG2BGmP37t1is8UK\nvCewXGy262XYsCeKqq0VOUAIdWrnzp0Tmy1aYL375Zi6Ag8KrBKTaajUqtVUvSBaCEJJKwrvoXRS\nOBITE91+bK174LZDrNYYOXToULa0LpdLWrToKGZzf4FVYjSOlfj4GpKcnHxJfjt27JDExERfViNf\n+EsrW7ZscY815gj86O4rXhI4J1ZrHZk/f74PW0FxJbysE6+RrSIul0sGDhwqFkusRETUllKlKsue\nPXsK1BivvfaahIU97HEnfVDCw2OLqq0VOUCIdWqLFi0Wmy1a7PZqAmU9ZglcEh5eTf7880+f2xQq\nhJpWFN5B6aTwfPPNErHbYyQysr5YrcXlo4/m5Jju0KFDYrGUEHBk9ZsRES1k+fLlPrb46vCXVkaP\nflpgnMdYY5PodNFiscTKgw+OEJfL5cNWUFyJvHQSVDvz6XQ6PvxwGs8+O4rz589To0YNLBZLgfIw\nmUzodJ6rZaRgMARVMyj8TPfu3ThyZC8//fQT998/ipQUJ9p/JQcuVzpGo9KTQqEIbG69tSuHD+9h\n//79lC9S6+YDAAAgAElEQVRfntjY2BzTmUwmRBxABmAFXIikKj93BcLCTBgMSTidF44kU6pULGvX\n/kCFChX8aZqigATyNjvuQX7RcuLECerUacq5c/fidNbAZnuVZ58dyJgxo4q8LIWGe+chb2nNKzrJ\nDy6Xi/btu7Jxo4XU1DuwWr+kWbMMVqxYgl4fVAvKBAyhqhVF0aJ04lu6d+/Djz+eIiWlL2bzD9Ss\nuZ8NG1ZiMpn8bdoV8ZdW9u/fT8OGrUhKehiXqyw224u8/fZ4Bg0a4CVTFIUhL534YqBsADYCR4Db\ngGjgf2hb6B0AegI5LQboNWd1+PBhXnzxNU6cOMOdd3ahb98+amtOL1IAR3U1WvFrp5aWlsaUKa+x\nefMOGjWqxZgxowr8lENxkXxqJeB8isK3hLJPCUQcDgevvfYma9duplatyowb9xTh4eH+Nitf+NOn\n7NmzhylT3uTcuST69r2DO+7oXmD7Fb7B3wPlkUATIAK4HXgFOOX+9ymgONrejpejnFWIUIBO7Wq0\nonQSQuRTK8qnXOMon6LIL8qnKPJDXjrx9vPhcsAtwCwPA24H5rg/zwGu+hbL6XTicrkKZaAiYPCq\nVryBiOBwOPxtxrVG0OlE4TeUVrxACPq9oNOJGvv4Fm8PlN8AngQ8f9GSaFvL4P63ZEEzzczMpH//\nh7BY7JjNdoYPH6VEE/x4RSve4tVX38BqjcRisXHTTXeQmJjob5OuFYJKJwq/orRSxEyb9h42WxRm\ns5X27bty9uxZf5tUFASNTjIyMujdeyBmsw2z2c4TTzyNmtH2Pt4cKN+Ktm/mZnJ/7HFVS7dMnDiZ\nL77Yi8NxAofjMB999Btvvz29EKYq/IzXtOINlixZwoQJ00lP34rTmcAvv0TwwAOP+tusa4Gg0onC\nryitFDHLly/nqaemkJa2EZcrmbVry9G374P+NquwBJVOxo6dyFdfHcfpPIXDcZD33/+RDz6Y6W+z\nQh5vru9yHdrji1sACxAJzEO7OysFHAdKo4k0RyZMmJD1uX379ln7mi9b9gspKU8CUQCkpDzGd999\nzmOPDS/6WigKzMqVK1m5cmVBLimUVnLTibf4+edVpKQMAioBkJ7+LCtWdPZqmaFKAbXiNZ+iCGxC\n3acEA6tW/UJqaj+gOgAZGeNZvbqxf43KgVD2Kd9/v4rU1BfQQqkjSEl5lO+++56HHhritTJDlavw\nKV6nHfCN+/OF4HjQguOn5HJNrgtDd+vWR/T6l7IW8jYaR8vAgY8U3crTiiKFgt2NF1QrPq/Pq6++\nJhbL3R4bjXwmdeq09LkdoUgBtFKkPkURXBRAJ1ejFX9XLyB59913xWq9xcPvfS2VKtXzt1lXpABa\nCXifctNNd4pONzVr7GMyjZChQx/3qQ2hSl46yU/ohR14Frgwv18N7XFFQblgxBTgRmA30JHcBZgr\nr7/+PMWKTcdu74nF0pqwsJlUqFCSpKSkqzBLUVQkJyczadIkBg8eDGhL4yxZsuRqsioyrXiDhx56\nkIoV92G3dyQsrCVG40D69u2mYsXyybWiE0XhUVoJHAYMGED16uew2VpgNDbGaOzFxIlP+tssIDh1\nkpiYyBtvvMHTTz/D8uXL83XN22+/RFTUq9jtvQgPv50SJZYyfnxOi3EofM3/od1ZbXN/twN/+qDc\nPEf/J06ckH797pewsDICz4nF0kNq1Gh8yf7zCt9y9913y5QpU6R27doiIpKUlCT169cPye1mExMT\npXbtZhIW1k7gebHb68ioUWP9YkuwkZtORNTWxIpLuZZ8SjCwZcsWsVqjRa9/VPT6kWK3x8rmzZv9\nbVbQ+ZSkpCSpVq2BWCx3i043QWy2eJkxY1a+rj127Jh8/PHHMm/ePDl37txVla/ITmF18of7380e\nx/w+UBYRCQ+PFdjmfgzhEru9s8ydO9cHTarIicaNG4uISMOGDbOOhWqn9ssvv0h4eG2BTLf+TorJ\nZJfz58/7xZ5gIjediARmp6bwH9eSTwkGeva8X3S6KVmP/uEdufnmu/1tVtD5lI8++khsNs8wlj8l\nMrJk4RtCcdXkpZP8hF6ko23wfoEq7mN+JS0tjdTUBKC8+4gOp7MCCQkJ/jTrmsZsNpOampr1/Z9/\n/sFsNvvRoqLn5MmTrFixgm3btqHXl+Xi+7AxGAxWkpOT/WleUHAt6ERRNCitBBZnziQgUsHjSAXO\nnfP/0pjBppOEhASczgpcXGijAikpCWzdulWF8AUpNwGrgJPAAuAg0MEH5eY68j958qRUrdpADIZS\nAncL7BP4Smy2WNm5c2e+7h6cTqekpqYW0b2IQkTk+++/l7Zt20psbKz07t1bypcvL8uXLw/IO/qr\nYcWKFRIeXkKioq4XiyVOzObiArME9ovROFpq124mTqfTZ/YEK7npRCQwZ38U/iPUfUow4XA4pEGD\n5gLlBf4Q+FMslrryzjvv+tu0oPEpaWlp4nA4ZPv27WKzxQp84x6/3ClGYzmx2eKlT59B4nK5vNFM\nijzISyf53cI6Fmjp/rwObWtHb+O2PTv33fcgn31mITNzEjAcWEJUlJ1Fi+bQocOVx/DvvvsBI0eO\nIjMzncaNW7N06f+Ii4srWuuvUU6dOsW6desAaNmyJbGxsQXZbvZqyFUnRVwIMTHlOHt2Nto7Hmcx\nGqvidDoRcWC3R7F69RIaNWrkdVtCgZx0AgXamvhq8IlWFEVLqPqUYOKvv/6iffuunDnzL9oKapGA\ng7g4M8eP/3Ph9/ArgexTEhISuPPOfqxcuQy93sAzz4yjTZsWPPjgKPbv34fL1RT4CtBjt7dh7txx\n3HnnnV4yWZETeekkP+JphzbSvpD2gip+KbRleZOrAJs0uYFNmx4EUgEDkEGnTkv58ceF2dJ+//33\njBnzEsnJKdx/f09at27OLbf0JSVlOVAZo3E01123k1WrlnqzLtcEq1atQqfTZT06uuA827VrB0He\nqaWmpmKzRaA9TDmAtsTmX8BvQC0Mhok0aLCSP/5Y5XVbgp3cdNK2bduA6NQUgUMo+5RgITMzk7Jl\nq3Hy5HNAb2AyMBVYQHj4IBISTvDWW9N4//1PCAsL4/nnn6B7d9/u+BzoPqV370EsWuQgPf114DF0\numXUrVudL7+cS506DcnI+BetX1mPXv8tL77YgjFj1GoWvqSwA+UlXBwcW4DmaC/4dSwK4/IgVwH2\n6TOATz9dDHQCMoFf6NWrKzExJTAaDQwZMoDatWuzdu1aOnXqTkrK+0AcNtsIrrsujpUr6+BwvOrO\n7TQWS2VSU897uTqhz6233prloNLS0li/fj1NmjRhxYoVEOSd2t69e6lWrRHQGW1ZzV+A54G9aA9c\nMtDr7TgcGQExuxLI5KaT5cuXB0SnpggcQtmnBAoul4uPP57Nb7/9QY0aFRkxYjgWiyXr/L59+6hf\nvwPJybvQ+lw92kTBMmrVqsbDDw9izJh3SUmZDiRisw3hq6/m0qlTJ5/VIdB9SpkyNTh27EvgcaAs\nMBSdbjkxMdOIiyvF9u01gKVAOeAYDRtWY9Om31Rf4kOKWifxwJdFlVke5BpL0rVrT4FJHm/e9hKD\nIdJ9bLzY7bGyZcsWGTr0cYGXPNKtl5iYymK3dxBwuI99J2XL1iiqMBeFB4cOHZI77rgjoGLErpb6\n9ZsLmAQyPPTUXuAz9+dfJTq6rE9sCTUu6EQksOIJFYFHKPmUQGHQoKFiszUXeEsslu7SrFl7yczM\nzDp/7tw5MZsj3P3rpRuOxMfXljp1rhP4ycMvTpM+fR7wY40Cz6c0atRWYJpAlMdKSSKRkZ3k3Xff\nFTAKlBCYIjBawK5W8PIxeenkarawPgLUumpZFQHHj58GmnkcOYjTORUYBEBycgT33DOQZs3qo9cb\ncbkupDtLTEw0MTEG/vrrOqAaLtf3zJnzmU/tv1YoV64cO3bs8LcZRcKOHfvRbjYTgWi0/1PnsFjG\nYjAsweVaxuzZH/nVxmAllHSi8C5KKwVDRPj8889Zu3YDVapUYPDgwZesBnH27Fnmzp1DZqYWe5yW\nNowdOxqyZs0a2rZtC0BUVBQvvzyZUaPG4HA8ysVJt8acO3eKuLhSwNmsPHW6M1it/l1xItB08v77\nr9KhQxdSUtKAZCAKEETOUbZsWbTtKWYAF0JW9Myb9z/69euXa57Hjx9nxoyZJCam0KNHN1q2bJlr\nWkXhyM9A+R2Pz3qgIRfXVvYLXbq0Zfv2V0lNbQlkotMdRKSER4o4du0SDh78BosF0tJMuFxx2Gyv\nMGnSm9x55x18++23nDlzhuuvn0iVKlX8VZWQYvjw4VmfXS4XW7ZsoUmTJuzcudOPVhWOXbt2sX37\ndoxGPZmZvdEWgXkALfRiL2+/PRWTyUSrVuOoUaOGf40NEnLTiUJxOaHoU3zJqFHP8MEHS0hO7oPV\n+i0LFnzFL798h9Godf2pqanodGFAuPsKPU5nFCkpKZfkM2LEUMxmAyNGTCIj436gAibTRK6/vh3D\nhw/kzjv7k5p6CJ0uAbt9OiNH+vZdjUD2KYcPH6Z37wcQsaDTZaLXd8DpHILZvIqKFY106dIFk8lM\nZqbnGKYkkZGHc83z2LFj1K/fgnPnuuJwlGL69G58/vlHdO3alQMHDrBlyxbKlClD8+bNvV/Ba4D8\nxGPc7/HZgRZx/qs3jLkM92x4dg4ePEitWs1ITT0L6NDrw3C5SgDzgAygL9o7iH2oWnU8Xbp0JDEx\nhb59e/g0bupaY/bs2VmfjUYjFStWpE2bNgERI3Y1fPDBLB5//BlMphYkJ6/G6awI3AZsAFYDdnr0\naM8XX/zPK+WHKrnpBALjDXVF4BBqPsWXJCUlER1dkszMg2jvUTgJD2/M11+/mbU6VEZGBjZbHE5n\nb+Bh4EfgObZs+Y0GDRpckp/D4aB27cbs2bMbcGAwRPDDDwvp2LEjq1ev5uOPP8NqDWPYsCHUquXb\nh86B7FNat+7M77+3xul8FjhFWFgDWrZsQMeOrRk16nHsdjujRz/Da68tQuRj4AxhYffz3Xef0rFj\nzq+CjR8/gcmTT+FwTHMf+ZYaNSbwyivj6N17EEZjS5zOv+nV61ZmznxbxTrnAy/rxGvkGktyzz0D\nxGh82h1n7BCoIjBcoJFAU4FHBO4U+EPKl6/r7dAWxRUgAGLECsqZM2fEYokS2OOOJzsiYBGoI3Cd\nwHcCT0jTpi29Uv61SjBqReF7lE6uzH///ede693pERPbWb7++uusNAcPHhSrtZRAP4HaArdIePj1\nl6S5wIcffig2W3v3exqZAvOlZs2mvqzSVeFvrYSHlxA4mvUb6HTj5Nlnx1+Sxul0yqRJU6RSpQZS\nq1YLWbhwYZ55PvroE5e9f7VFypSpKXZ7tMDv7mMJYrdXk1WrVhWq/a4V8tJJXqEXf+UlDqD+1Srr\nakhNTcVgMBAWFsaePQdxOHqhLQ0HUNn970q0R0gPAFHYbI/Rq9cdvjTzmqRevXq5ngvWO9ljx45h\nMpUkLa2q+8iFOLIngR5oD1dG0bmzb5dBCmaupJOtW7f60BpFIBOKPsXXxMbGUqtWbbZtG0lm5jBg\nJXr9Vlq1apWVJi4uDpFU4EGgNXACl6sRlStXzpbfgQMHSUm5Hm3YkAS05d9/H/dJXXIjGHxK+fKV\n2b79e7SH8+nYbCuoWnXIJWnS0tIYNuxBxo59Er3+yhsm9+hxO7Nm9SIlpTlQGpttBD16dOW9997l\n4vtbEeh0TTh06FDRVugaJC+PU/EK1x4oOjNyRESE5ORk7rrrPn78cSk6HTzyyKPodDpmzNhLauoC\nwInZ3IaMjN2IZAImDAY90dGx3Hdfb6ZMmZgVj6XwDgcOHMjzfKVKlSDIHpOmpKRQunRlEhI+Am4B\nVqDTdUckBe0+0URkZAzHju3GZrMVefmhyJV0UrFiRb8/JlUEBqHoU/zBqVOn6N//Edav30D58hWY\nPfudSwaXCxd+SZ8+95ORkYZOF0lYGDz11EgmThybLa8lS5bQs+cwUlPD0N7pN1CnTk3+/nuD7yp0\nGcHgU/7880/at78ZkRo4HEdo374RX331KQaDgdOnT3Pjjd3ZvFnbKMVgCOODD6YxaNCAKxb8+edf\nMHr0JFJTU+jT5y5efvl5qlSpx+HDT6ItbLATm609v//+E3Xr1i2CqoY2QR168cADw8Vi6SmQJnBK\nbLYm8u6770nnzndIWFiEGI02MRojBRa7HzcsE7s9Vs6cOeOX6XtFdgjSx6S//vqrFCtWWiyWEmIw\nRAjcnqVDaCJGY0N57LGnvFb+tUiwakXhW5ROCs+ePXvEYokR2Ohe8m2alCxZOcftkx0Oh2RmZkp4\neEmBD9zpfxerNUb279/ve+MLgL+04nK5spbZO336tPzwww/y+++/X9K+2lK3g90hpEcEqojRGCUb\nNmy4qrpu27ZNSpeuIhZLrJjNEfLxx3OuKp9rkcLqpBXa20vJaLt7uICEwuvrioiISNWqTQTWecTi\nvC81azYVl8slx48fl5tuuk0g3uO8iE5XU1asWJHvBkpNTZXNmzcH/H/4QGfNmjXStGlTsdlsYjQa\nRafTSURERFB3ag6HQ9auXSs6XbTAWg+dfSDQVerVa+PV8kOR3HQiogZAiksJRZ8SKMyYMUPgpsv6\nTpucPn06K016err07j1QDIYwMRotYjBEXZIe2kr16vXlv//+82NNAs+nvPLKVDGbw8VgMMlNN90h\nCQkJ2dKcPn1awsPjBA54tOdEgfry2GOPXXVbOJ1OOXbsmKSlpV11HtcieenkysEwMA3oA+xG25lv\nEPBukUgsH8THlwHWuL8J8Au7dv3DrFkfMW3aDFatOo22tu2/7jT/IfIv58/nb6e9PXv2UKlSHdq2\nvZdatZoxaNCwrG0wFQVj2LBhLFiwgOrVq5OWlsaHH37II4884m+zCoXBYOCPP/5AxIa2XTVoOlwH\npBAfX9p/xgUpoagThXdQWrl65syZR8OG7WjUqD3/93+fZzt//PhxYDvaHBjA34g4LkkzduxEFi8+\nitN5EodjB05nKnBhab5E4DD//FOLu+++cqiANwkknSxZsoQJE6aTnr4Vp/M8q1bZGTx4xCVpFi/+\nivj4aiQlOYC17qMu4HfgFNu378lKu2zZMpo160S9em14++3pVxyf6PV6SpUqdcl62Qrvc2HNZM+o\n+C0+KFdERHbs2CF6fbjADQKtBaoL1JQ6dZpJ06Y3CCwTeE2grMBdAiXFaIyVv/76K193EY0btxWd\n7g333dx5sdsbyueff+6925YQpnHjxiIiUq9evaxjDRo0CPrZnw8//ND91CJaoJNAM9F2UbJKly5d\nvV5+qJGbTkTUjLLiUkLVp3ib+fM/FZutksC3At+IzVYu20oWS5YsEYMhXqCGQC+BONHpTJKampqV\npm7d1gIr3P3jZx5+sLtANYGhAmfEbA73dRUvIZB8ymOPjRKY7DFLvEvi4ipnnT937pxYrcUF1ruf\nlhcT6CxQXyBWoJaMHDlaRERWr14tVmucwOcCP4nNVkfefPMdbzblNUteOsnPjHIyYAb+BF4BRuLD\ngOeaNWsSFxeHtg7kE8BmYDaHDh2nbNmS6PV/uI9/BSSj1xvo0KE1derUyVf+u3ZtQ6S3+1skKSld\n+fvvbV6oSehjt9tJT0+nQYMGjB49mqlTp4bE7Pxdd92FwXAa7a3we4EOwB1AaX79dZ1fbQtGQlUn\niqJHaeXqePfdeaSkvArcDNxKSsoLvPfeJ5ek6dSpEzVrlsZkKgOYsFiiGT78MSwWS1aaMmVKotNt\ndn/bDvRH6293oe0k9w6wmeLFS3q/UnkQSDopU6YkZvMmLo67NhEXd7F9Dh48iNFYGm11ihZoC4xt\nQRtWtcFqPcGwYQ8B8PHHn5KaOhq4C7iBlJR3ee+9eb6rjALI30C5nzvdMCAFKIe2PpbP6Ny5A1Ae\nbXBiAxIoUSKW119/nqio6djtPQkLG4/Fsp6RI++lffvmjB8/gU2bNl0x7ypVaqLTLXR/S8ZmW0at\nWjW9V5kQZt68ebhcLqZNm4bNZuPIkSMsXLjwyhcGOJGRkXz//dfACmAOcBRYhNa5qBUvCkqo6kRR\n9CitXB1mswktNOICiRw79i+ZmZkeacz8/vtyxo/vxIAB4XzwwdO8+ebLl+Tz1lsvEhn5CjZbL8zm\nb9DpvkTbmKQ0MBqdrg822z189NE7+JNA0snDDz9ExYr/EB5+Izbbfdjtw5kx4/Ws8+XLl8fhOMrF\nB/PJQDIGwwnKl9/O1q2/X1jVBbPZhE7n+TsmYDKZ8iz//PnzTJ06lWeeeZZff/XF3nAK0AbF/gh2\nyZoS37Nnj9jtsQLjBaaL1VpWFiz4VERETpw4IR999JHMmTNHtm/fLiVKlBeT6QHR6caKzVZCfvjh\nhzyn23fs2CElSlSQyMhGYrWWknvvfSDHt34VV+aLL77I8QUCQuQx6ZNPPikQJlDa/fixmEyd+obP\nyg8VctOJSOhoRVE0hLpP8RYrVqwQm62EwFSBVwSixGxuJtdf30UcDkeB8jp69Kh89NFHMnv2bOnR\no69YraUlIqKBREaWkIkTJ8qOHTu8VIv8E2g+JSUlRT799FOZOXOm7Nu3L9v5//3vc7FaoyUqqqmY\nzcWkb9/+smjRomy/zY4dO8RujxWd7nmBaWK1lpYvv/wy13Y4f/68VKxYW8zm3gLjxWotLfPnLyhA\nS167FFYns4FDaPtD30rem5QUJZdUYufOnTJ48DDp0eM+iY+vITpdhBgM0TJq1Cg5c+aMZGRkyNix\nz4rRONQjNmiR1KnTKteGSUlJkYSEBElKSpLff/9ddu7cqQbJhaB///4SHx8vffv2lW+++SZraRxC\npFMrVaqyQBmBSAGbQDHp1auPz8oPFXLTiUjoaEVRNIS6T/EmP/30k+h04QL3ibYEXKaEh9ct0IpQ\nl5OZmSnr1q2Tr7/+Wjp2vE0slhgxGIpJXFwVmTv3k6IzvoAEuk+ZMWOWlC1bU+LiKsuYMePF6XTK\niRMnZM2aNXL06NFcr9uwYYPUqtVYTKbiEhtbWebOnZtnOdOnTxertYfHGGitxMVVyl8jXuMUhU7C\ngG7AfLRB84eFzTAf5FiZWrWaCrQX2C2wQCBCDAabhIXZpE2bG9x3zxdEsinbFtYul0s+/niOe9Bj\nF4iU+vWbS1JSki9+i5AnPT1dFi9eLH369JH4+HgZOHBgQDiqwjJt2nSBcgJLBeoKmAVMotPZ8/3i\nqOIiOelEJDA6NUVgEao+xducOHHCvYW1K6tPjIzskuP21J4cPnxYBg58RG6+uad88MHMrMmjOXPm\nisEQ7p4oMAqY3D7xT4HVYrWWveITXG8SqD5l8eLFYrNVFG150b/FZmshkyZNkfXr10vPnvdLt273\nygsvvCDdut0r99wzQP744w85duyYlC1bxf0E0yxQWnS6J6RcueqXvGx5OZMnTxajcaTHGOiohIfH\nFlkbhzJFpZMw4Da04MzTRZHhFcixMjpdpMAutwhaCzzvdgR7xWwuJWZzCYHVArvFZusgTzwx9pLr\nX375dTEaY9yD5FcFZgmUkPbtO/vit7gmSE9Pl6+//lq6d+8u0dHRfndUheX8+fNiMpUQ+ESgncAE\nt+b2CcTKsGHDvG5DKHK5TkT836kpApNQ8ym+wOVySb16LcVofFK0tXrnSFRUqTzXPP7vv/8kNjZe\nDIYxAp+I3d5Qnnlmgmzbtk10ugiB5gJtBRIFUgRuFnjW3R+/LkOGDPdhDbMTiD7lnnsGCLzvMXhd\nKZUqNRSbLVbgTdHW5C8mMEzgdbHbY6VWrSYCxQX+cl8zTaCyRETUki1btuRa1ubNm8VqLSHwvcA+\nsVh6yD33DCiy9g1lCquTW9DCLw6ivcl0C74Jv8ixMgZDtMBP7oGKUbSd0jQBms2PSN++/aRs2ZoS\nG1tBhg0bJRkZGZdcHx19YTmcSR7CXSJhYSV98VuENEuXLpX+/ftL+fLl5b777pOlS5dKZmam3x1V\nYfnss88Eyou25I9ZIMlDO0Pk7rvv9roNoURuOhHxf6emCCxC1af4iuPHj8sNN3ST4sXLSt26rWTT\npk15pn///ffFau3l4d8OiNVaTMaNGyfQUKCbaEuVXTj/rcCNAiJ6/aPy5JNP+6hmlxLIPuXBBx8V\nvf4ZjzabJzExVQRe9zj2eVY7wmTR6cIE7vE47xIwi9kcK3v37r1iW1SsWE+io+Olb9/BkpycXCRt\nHOoUViefAt3RNhvxJTlWZsyYMe5HP+Pcd1w/uYWULnZ70zwD3UVEihUrI9BItLWXL4jwJ7Hby3mj\n7a8pevXqJYsWLcr2aIgg7tRcLpfccEMXgVYCcaKtn7zMrZsMgboyZ47aJrQg5KYTkeDWiqLoCUWf\nEqi4XC6ZNm2aWCz3e/SNx8VsjpDXX39doKlo6yY/5nF+tEATgYESHV1Ojhw54hfbA9mn7N+/X6Ki\nSonROFR0ujFis8VKhw43C0z3aMdvBDq4B8Q3iPYOTBWBZPf5PwTM0rv3wKxQGJfLpd6pKkK8rBOv\nkWuFrruug3uwO8A9aO4kUEFKlaoqO3bsyFM8o0ePk7CwGgIRAvPcAi0rzz//gjfaXiH+d1SFYcmS\nJWKzVRdtQ5uebr1Z3ZqrKC1bdizwW+SK3AlmrSh8h9JJ0eF0OmXIkOFiMtncf8VEWy3jR7HZ2smD\nD46QhIQECQsrLlp4QAWBlgKtRKcLl7Zt28vEic/n+VKaPwkErRw6dEgmTXpBxo0bL1u3bpUVK1aI\n1VpS4FOBrwRKCTwu8KT76eW3ooVjxAt0Fp0uXHr16i2JiYkiIjJlymtitRYTo9EsPXr0k5SUFG82\n4TWBl3XiNXKtUOPGHUWLwZkhUEfgPYEvBVqK0Vhc2rfvmmvAu9PplJdeelWqVKkndnsZKVeurrz5\n5tveanuFBIajulpeffVVMRpHCNzqHiB/J9qLpH0lNrbcJW9XKwpPMGtF4TuUToqOIUMeES2s4rjA\ncWHlZQQAACAASURBVDGZmkr16g2lUaP28txzL2T5uBEjHhOo5e5zbxB4T+z2GD9bf2UCVSvLli2T\n1q1vlubNb5QHHhgizZp1kkqVanms3HXI/eTcICZTuERE1JBixUrLyy+/7J682StwXiyW7jJ48KNF\n2GLXJnnpxFdLvRWK5ORkZsyYwb//HqdDh7Y0alSbbds+Jz09EXgKbU8UgEgcjudZty6MSZOm8OKL\nE7LlpdfrKV26BMeOHSc9PYkyZWowYEB/H9ZGEUzUrl0bvX4qYEfb9OZbtF3d/+bUKWH+/Pn076/0\no1Aogo+MjAxmzfoU6Aq8CdxKZuYzREZOY8OGn7LS7du3j/nzF6NttlQaeAG9fh01a9b1i93BzunT\np/njj020atWA2267mSZNmvDbb78xePDjOBwzgbXAXCABCCMzcyCZmUOAXYwfP4T09GeBKgCkpT3H\n99/3y7UsReHJz858oG2HV8ObhuRGWloaLVp0ZOzYX3n99Sh69hxBfHwpatTYhtG4HG1L6wtsBeJI\nS+vF+vV/5ZjfypUrGTDgUVJSFuN0Hmb9+pK0aNHRF1W5JkhJSWHXrl3+NqPIuPnmm6latQyQivZf\noBRwAHgG+JohQ0by999/+9HC4CTUdKLwHkor3mP79u24XBloc2ZWtK2SF2CxGLLSOBwO6tZtwalT\nDwIngClAR2Ji3uCzz2b5w+wcCRadnDlzhvr1WzJx4h5ee83GjTfeTVRULJ079+DQoaFobfwg0BKY\nCQxBm6hpD5QjIyMJk+lPjxy3EhdXwtfVUFzG7Wgbux9wf28EfO2DckVE5P/+7/8kPLy9XFwLcr+E\nhdklMzNTfvrpJ4mOLicGw23u+NFo0da4jZRy5WrJiRMnsk2vt2rVSmC4RxD9SQGzr2f5Q5KvvvpK\nqlevLhUqVBARkf9n77zDqyq2KP67vaVQUiAECITemyBVULCgIgpWUBSfCIINsT0L2LGAvaIgikpT\nUdEnikpRkY4giEpVpEpNSL93vT/mJCRKCZBu1vflS3LuOXOm7Ltnz549ay9btkznn39+id36yi8G\nDBgg6JRLBn+x4sdCgqs0evToQq9DWcKR5EQquduk5SgelFWdUpQIBoPavn37PxigJOmhhx4SXJFr\nPvxaEKEvv/wy557vvvvOmluV66eVbr755qJsxlFRmnTK2LFj5fH0z9WX862zLwl5+tjprC8YnOva\na4LOioiIVo0aDRQInCef7xqFhUVr0aJF+X7/vn37tG/fvuOud1nH0eQkPx7lUUA7YK/1/3Kg9kmL\nVz6RnJyMVA2wWVeqkpWVgSTOOOMMVq78gSFDahIZ+RWQDtwHrGHLlm7UqtWcffv25SnvwIEDwFoO\n9claSkkESonHqFGjWLhwIRUrVgSgZcuWbNiwoZhrdfJo1aoVEM8hGYwHUoAQsJaMjIziqlqpRFmV\nk3IUPMpl5eSwYsUKqlZNJCGhMZGR0UybNj3P56mp6eSdzuPx+Tx0794958qePXuAg8BO60oasIPx\n4yfz5Zdfsnt3UaRVODpKk5wkJSWTmRmf60o8hlQsCdhvXTtAKLSDv4+N3b6GGTOm8NNPi3jxxb6M\nHduWlSsXcsoppxzzvRkZGVx4YT+io6sRHV2N3r0vL5+78on8GMqZwL6/XQsVQl0Oi27dugGzgCnA\nOjyewXTtejYul4udO3fSufPZTJz4A2lpAaAjcAlQDXielJS9XHXV4DzlXX755cASTEzWcKA3Ho+r\nqJpTpuFyuahQoUKea3Z7fqN7Si6+/XYRZhNlOrAOGADUAXoAa+jfv38x1q70oazKSTkKHuWycuII\nBoP06HEBO3c+QlraX6SmzuGqqwbnMSB79ToPv38c8AXwC17vMK666oo85XTp0gXjJGgD3A50BrqR\nlJREnz4jSUxswpIlS4qsXYdDaZKTc8/tidc7Hvgcs1k/BBPyciXQFriFQKATXbq0x+9/DVgIrMDr\nvYMHHxxBt27dCA8PZ8CAAQwePJhatWrl670PPTSaWbP2kpm5i8zMXXzxRRKjRj2ar2fXrVtHx45n\nERubSPfuvdm6desJtLxsYzzQD1gF1AWeB17Jx3NeskcY1gCPWdcrAV8Cv2K+nRUO+3SuLY3vv/9e\njRufqgoV4tWt23natm2bJOmKK/4jl+tWawv8U+s0bpa1TbFF4FUgUDWPez0YDKpdu26CSEE1OZ1h\nmj9/flF6+MssrrnmGk2aNElNmjTRr7/+qmHDhun666/Pz9bXychKobZp+fLl8npjra2xNtb22LkW\nA4ZPdnvYUTNdleOfOJKcSPnaJi2xslKOgkdZ1ClFhU2bNslkoF0qyLTmxa56880389z30UcfqXbt\nFoqJqa0hQ25Venr6P8oyjBexgkcEUwRbLR2YLHhPtWs3K6pmHRalTad8+umnVvK0CoKeMonTQoLz\n1KBBE3344YcWr/XLio9vqKpV6+mhh0Yflvo2FArp559/1vfff68DBw4c8Z2dOp0rmJErlONjdehw\nzjHrmpSUpJiYBNntYwS/yOm8R3XqNC9zjE/5kJOjIgA8inHDLgEeIf/JR/zWbyfwA9AJeAK4w7p+\nJ+ZkwOGQ04CMjAydeWZv+f01FRHRQvHx9bR582a1bp1NEydLEbQQdJZJqVlH0Ftud8w/OiQtLU2P\nPPKIbrzxRn333XdFORZlGsnJybr77rvVunVrtW7dWv/973+VmpqaXwE8UVkp1Da9//77iojoJUO0\n39SSraaW4XybfL7IMqcwChtHkhMp38qqRMpKOQoeZVGnFAX27NmjunVbyPDw1rHOWPwuiNHYsWOP\nu7ysrCw1bnyKnM5Olg6sJXjAmnv3ye0OFEIr8o/SplOCwaBsNrvgK0GUDH/y1fL5Kh82aUt6erqm\nTJmil19+WYsWLdJbb72l1157TZs3b1b//tfJ54tTREQbVa4cr59++umw77zmmiGWYzE7Bvo2XXXV\n9ces67x58xQR0TaXgR1SIJCgX3755YTaXlKRTznJFxxA5Ak85wcWA40xQcGx1vUq1v+HQ04Dnn32\nOfl8PQTpAsnheFBnnHGBbrhhuLzeyy0jOU12e3ZGmxiZQ30VFBdXL09npKWlqVWrzgoL6yKf71r5\n/VH67LPPinpMyjyysrJyDgxwfAJ4vLJSqO1Yu3atnM4KgjmWd2aIYJJghcCrN94YX6jvL+vILSdS\n6ZaVchQuyopOKQoMHDhUbvcgy0sZlEmHXFk2WxV99dVX/7h/3759WrZsWZ4D8Hv27NF//3u/rrrq\nel16aT95vfFyOrvIbo+U3R4l2CGTunqMmjfvWJTNOypKi06pW7eFbLaXBSsFI+R0VtCMGTP+cV96\nerpatOgolytBdnsjmVTWZ8nvv1Jeb4R8vhaWZ1+y2capceN2h33fjh07VKNGA4WHd1F4+GmqXr2+\ntm/ffsx6Ll26VIFA7Rz7Cw7I46moP//884TbXhJxnHLyD7wHRGA8y2uAPzm0yjoW7JjtjCTM6gwO\nHQoEE/i09+8PWchpwLXXDhU8k2tFs1JVq9bT+PHjVb9+K3k80bLZwmWzVZQ5nRsjGChoqUAgNs92\nxVNPPSW3u4ngfWu7Y7aqVq1THONS5nDZZZdp//79Sk5OVsOGDRUXF6fHH388vwJ4orJSqG2aMWOG\ntcXosjwzb1uykyqvt55WrFhRqO8viziSnEj5VlYlUlbKUfAoizqlKHDKKd1lkiNlz5nTBNVlt1fS\n3Llz89w7e/ZshYVFKyKiqbzeCnrxxVeVlJSkhIRGcrsHCl6wPMgjBBmCybLbI+R0BhQI1FB8fH1N\nnz5dEyZM0Pfff18s7S1NOuX777/XhAkTNHnyZFWuXF0QLfDK4QjXvfc++I/7X331VdlsUYIB1ljU\nt7z6ElxgjcshFi+fr8IR352cnKyZM2dq5syZSk5O/sfn69ev18SJE/Xxxx/n7JQGg0GdeeaF8vtP\nFzwuv79tvjzRpQ35lJMjIpuwrx8wBnBh4pWPB5GY7Yxu/FPY9hzhmZwGvPjiS/L7uwhSBCE5nXfK\n5aqssLBeCgTOltdbWS7XJTLxySky6a3by2TsC9PmzZslSUuWLJHHU8ESrs6CtoLN8nojimtsyhSa\nNTNxapMmTdLw4cOVkZGhJk2aHK8AHq+sFFp7XnjhZTmdkdYk0ccymHsLuggay+OJ1N69ewvt/WUV\nR5IT6biVVYmRlXIUDsqaTikqXH/9zXK7r7LmxExLb52j8PBz9OGHH+bcl56ervDwaBlaOAnWy+eL\n1pgxY+T3dxeMlkmt/Ial/7rKpK/uJY8nUlOmTNHIkY/I76+uQKC//P6auueeB4q8vaVFp9xzzwPy\n+2tafVVdYWFVLCfgHsEOeb1VdPHFl+u++0bmpATv37+/8tKTbhd4rbF9QjZbomCv5VF+Ri1adDqh\nPpw9e7b8/iiFhV2usLA26tTprBxKwczMTL388su68cbhevPNNxUMBk/oHSUZR5OT/PCiOTHGcW/g\nRQwLxvFa3vuBT4HWGDbtKsB2TIqfnUd6aNSoUQCEQiFat3azZEkCDkcEUhKpqQNJTs5e+NUDrsUs\n+q7AOMDPAyZiszlJS0sD4D//uZX09GcwrAUC+mCznUm7dh2PsznlOByysrLIzMxk3LhxJCQk8Mgj\nj7Bz5xGH90g4blnJlhOArl270rVr1+Ov/N+QnJzMbbfdQVZWP2AahhXxeYycCbiIHj1q/+OkdTmO\njWw5mTFjBh07dsyRk9zjmE+UCFkpR+GhLOmUosSTTz7E0qXns3RpVSQH0BS4mVDoatq0eS3nvu3b\ntxMMujB2IRg6smasWLGC1NSfgCigFSY8NxNDizkcaEl6+mruuONeduzYRlraGky37mLMmIZcd90A\natasWWTtLQ06ZfPmzYwZ85zVVzFWsbWBPkBF4APS0jKYNq0JTuefvPxyO1atWkRiYiLGRrdhssKu\nBYLAPjye+dSpE8W6dYm43TGEhWUyder/jrfNAAwYcAMpKW8DZwNBli8/g/fee4+rrroKp9PJ4MGD\nj1VEqcKcOXOYM2dOgZV3Eybc4n+Y7YkEYH4+novi0ClRHzAPOAOzrXGndf0u8nGYTzInOzds2KDV\nq1crLq6+oIr1c4fgLME1Mqd7a+WKpUmSzRamTZs2SZJiY+sIfs61TfG4oIYqVqym3377rTgWMWUK\nzz77rOLi4nT22WcrGAxq48aN6tSpU35W9CcjK4XSlk2bNsnvr2btPvQXNBCsyiU7Y3T11YMK5d1l\nHUeSEylf3p8SJyvlKDyUJZ1S1AgGg/riiy9Uu3ZTgU1RUTU0e/bsPJ8PH36nzNmLKoInBZsEFeRy\n+S3PcbYXc5NM+FlVwYVWuMAL8vsrKiKiaS69KEVEtNYPP/xQpG0tDTrlhx9+UERE6zx9BTUtj30b\nGSau2TmfOZ3X68EHH9LWrVvl81WSYVuqZtk74QKnnM6KqlgxXvfcM1KrVq1SWlraCfehxxNuebYP\nHfb7NyXTyoecHBds5M8T3RRYhon7WYkhYARDuTKb46CHy4333psslytBJgB+g6CDDGVXbevL3iCX\nEIbk99fU2rVrJUkXXzxALtdVMnFWf1r3vi94XN279y6SwcjIyND99z+kzp3P0zXXDMlXMH1pRSgU\nUmZmZn4E8GRkpVDqnpmZqZiYBEsx3S24VnCltQjbKkhUhw5dC+Xd/zZky4mUL2VV4mSlHEWH0qxT\nihOH2yp/7LGn5Pe3E6wTrLacTGGC5wT/kQk3y55LUwQOwV/W/6sEPnXrdp4iI6sIPrSM6pmKiIgt\n9sxvJVGn7N27V+HhMYJPrL76QIY9ySd407JhVufq81EaMeIuSeasjN1eKVf//yITfvGjYJX8/vqa\nNOldScbGGDny4eO2MTp3PkdO5wiZw5+/ye+P17x5846z50svjiYn+TF4R1oFZKclyy7swWM8twqz\nZ/N37AG6H+Z6vjBt2mdkZt6LkW0wbHXXAquBu/B43iYz8wlCoV44nW8THu6gR4+LSEtLpU+fXsTH\nf8XGjX6Mc3wkcBGwkE2bpp5olY4L/fr9h5kzt5OaOoQFC+Yze/ZprFmzhLCwsCJ5f2HigQcewGaz\nYWQObDbbMZ7IQaHIysnA6XTy1Vef0KxZJ6RxwEsY+vAARnZuZNWqScVRtVKPI8nJ/fffn5/HS5ys\nlKPwUJZ0SnHicMk3pk37lJSUB4FE68qDmDCzG4BTMMmV3sF0471AdaCydW8TwElCQjQdOlzL009f\nR0rKxdjtXoYMGUZk5ImQY504SotOadGiOd9+exlSBg6Hj2CwP8Z2GQD8DAwFngO24Pe/TJ8+Mzh4\n8CBfffUVbncj0tKy+78eZizCgNqkpNzB88+/yubNG5k58ytWrHCQmnoDCxZ8y5dfdmHNmiWEh4cf\ntW7Tpk2gZ89L+PFHHw6HiyefHEPnzp0LsvmlFvkxlA9yyDj2YYJ/1xRajY6BtLQk8rK0rMNkSUvF\nZpvNffcNZ9asOfz662vEx8ezevVB0tImADG89dZgOnasyR9/QFaWH6MQMoExdO7crtDrnpyczIcf\nTiMr6y/AT1ZWb/btW8Y333zD+eefX+jvL2wEAoEcBZWamsrMmTNp1KhRMdfqxNGkSRN69+7Jhx9u\nwuy8JQE3YmjFfyAiYmZxVq/UoqzJSTkKD+WyUnioXLkCZv4807ryC+Ys2ypgN8aJejsmlnYXxkGw\nHGgJTAQymTBhCTbbfox9Op9QKMTzz19Bhw7t6dWrV5G1pTTIyQUXXMHChQlILwPzsdluxuVaTWbm\nFiAdeBgYAXSidu3aPPPMOJo2bUrLlp34448o0tJWYvKibMI4bZKBVKv08Sxdup3Fi3cRCs3HnEMM\nkJXVm/37l/P1119zwQUXHLV+sbGxLF06l9TUVDweT4nNbFha4AHmFsF7Duser1gxzoqpGiCTBMJn\nbQl55XTW1ttvv51z76BBQ2XYLWrIUKo8ovj4xqpTp7kcjvoCt8Cjhg3bKCkpqdBd+0lJSXI6fdY2\nltleCQ8/XR999FGhv7s4kJaWpi5duhR47E9+5KSg8Oabb1ry5pPJolRVJpNSuJ577rlCffe/Bdly\nIhV8nFhRyko5Ch9lQaeUFCxfvlxhYdFyOofJ5bpOXm8lOZ0eOZ0B2Wzx1ha8BFlyueLkcsVY2/3h\nlk48VYZT/ixLN8ZZ4RpP6tJLBxZr20qaTklLS5PD4bbCPs3cHwhcrKZN21nhLs0Et8jhiFFYWJxa\nt+6qZcuW6aWXXpLP1ysnrMX0e5wVqjFSECmn83yZ/BF7ZPiUvYKDuWyM7nmYTspxeBxNTk5kyRAA\nqp2wWJ0k9u9PwWwPNQPiMSdG2wK7cTp7sG/fvpx758//AeM0/wR4CHiEUCiNFSu+47nnbuK+++5k\n9uxPWbNmcZGEPoSFhXH++Rfh8/UBPsHpvJOIiD/o1q3bMZ8tjTh48CB//vlncVfjpBAKhYAQUAOY\nBbyNOXmcSvv27YuzamUGZUFOylE0KJeVgkOLFi1YsWIBDz8cz6OP1mPdupUkJ+/nzz830KxZDTye\n/wAz8XiuoVmzOkyd+ioJCfWABzAe0M8xpBCrgCnARxiv9MO43cXrjSxpcuJ0OrHbHRjvPICw2bZy\n773DadGiOR5POjCZYLALycmzWLr0Sk477Ww2bdpERkZ9TOTruZj+/ggTqjEKu70/7drtJRCohmHO\nCAB9rXs/wem8i/DwTZxxxhlF3eQyhfyEXuTmTLZjeE2OFZ9caHC73aSlVQPOsa7cDmQAP2CzvU+3\nbjfm3Pvbb79g4vCrYgzreaSnf8Dq1at5++0Z7N69l7Q00aVLF1wuV573SOKJJ8YyYcJUAgE/jz12\nN2eeeSYni8mTx/PAA48xZ87L1K4dz5NPzjtm7NDfEQwGeeih0bz33kdERkbw1FP306VLl5Ou28mi\nadOmOX+HQiF27tzJ/fffz0033VSMtfon3nprEk8++QqSuP32wQwYcOUR723Tpg1mE+UlzIIMYBRw\nJ1WrVi30upZFHElOylGOv6O06JTSisTERO68884812JiYpg373/ceedIfvzxRVq1asSddz7CnXc+\nQHJyCnb7E4RCAtzAdMwcnD03vgpczsaN62natBOhUIhGjRJYtWo9gYCf0aP/S48ePQq8HSVdpzgc\nDh544EEefvh0UlIG4PUuJjFRpKam8ttvLtLTl2POCr6D6dcmSLPw+/243S+TmnoBsBUTAuPNVbKf\n0047jQ0bJpOS8izSOcBmYBUu10Dcbg+Jic3YuXPncdsZ5Tg+JFg/NTGeZNfRbi5AHNY9fs89o+Rw\nNBJ8Lhgnmy1MHk+4qlato5kzZ+a51+OpJFie6xTpxQoPj1IgEGVtXXwvn+90XXfdjTnPhEIhbdy4\nUbfeert8vpaCuYLp8vtjii3r0N8xYsQ98vs7CL4VvCu/P0o//vhjcVdLGzdu1MaNG7Vp0yZt2bIl\nh6ycErD1lY0pU6bK70+w5Odz+f0JmjJl6hHvX7RokaCi4L1ccnS/IEwHDx48qf76t+JIciKVLFkp\nR/GjNOiUsoDk5GT99NNP2rNnzz8+CwaDatWqs5USe6FstlFWCEBXwcWCO3Ppxg8EteRwxFk6dqCg\nkWCeYJp8vmgtXLiwwOtfWnTKzJkzdfvtd+nZZ59VSkqKnn76abndw2SSh3hlGJUMY1dYWFdNnz5d\nU6dOs0JOwwSXC5oLZgleUyAQpTVr1ujXX39Vs2YdZLOFC66XYQO7QrBQdvtoRUXVKDHJsbZt26Y1\na9acFJVdYaAg5KQF5hTTMKD5yRaWTxy2MaFQSE899Yxateqmrl3PPypf4+DBQ2XSWT8pQ3dTUa1a\ntZPLNTTXF3uLAoHKkkyWoh49LpDPF2uljGwl2Gfd96iGDr21UAboeBEVlSBYk9MGm+0u3XffSIVC\nIT377Atq2bKrOnfuqfnz5xd53ZYvX67nnntOzz//fE56Z0qQourW7YK/Gb2T1a3bBYe9d/Hixdai\nqrI1MTwkw9sdI4cjUd9+++1J99e/FYeTE6lkyUo5SgZKuk4prdi/f7+uueYG1arVVE5nhMLC6snr\njdS4cePz3Ld582b5fLG5YpaliIgO6tPnMjVocIpstoAMF/AjgmjZbLkdC40Fi3Pp24d10023FUp7\nSqNOWbZsmXy+GMEimbTUiYIxstkuUnh4XE5W4ZtvHiF40IpVflrQVh5PbJ5Fx88//6xAIEGwS4Z2\nLjPXePXQxx9/fFJ1LQjcfvu98ngqKDy8rqpUqZ1D3VsScDQ5yU8g0c3AJCAaiLX+LrZ9L5vNxm23\n3czSpV/zzTcf067dkdkqRo9+hEsvPQOn80Hs9rc4//we9Ot3CXZ77syU+3A4nPzxxx889tiTfPtt\nkNTU35F2YNYE/wXAbt+Hz+cu1LblF263B5NAyMDh2IfH4+bJJ5/m7rtfZfnyu5g//xLOOutCli9f\nXmT1evbZZ+nfvz+7du1ix44d9O/fn+eee67I3p8f+Hx5+w72WdfyQhLdu/fi4MGXgb+AycDjmOym\nc7HZMnG7S4Y8lDaUBjkpR8lAuawUDkKhEGec0Yt33klm48atZGVNJzn5F9LSFnHTTXeybt26nHs9\nHg/BYBqHGBbSyMraxfXXD2Tu3JnMmfMZjRrNwWZ7GLc7lcTEWkD2WaG8+tbMo//UtyeL0ionLVu2\nZMKE54mM7AU8itv9F3b7RKRKpKZewBln9CIlJYX09BTs9n2YWOVbgEeJj4+jbdu2OWWZcUrB2HuZ\nGMIyACHtK/b5atasWbz00lTS038jKelXduy4jT59BhRrnQoSqzAR4tkIkDduubBwVOs/KSlJn332\nmWbNmqXU1NR/fD5q1KNyu8Pk98epRo0GWrdunSRpw4YNstvDZBgzXhBUkt3uk9cbo/Dw6jKneLNX\nv98IGspmu0fh4TFav359YSxkjhuvvz5efn9NwQtyOG5XxYpx+vPPP1WjRhPBwlz1H6Vbb729yOrV\npEkTJScn5/yfnJysJk2alKgV/ffffy+/P0owWjBafn/UYUNq5syZI5OJKncWpYstz3JVORwR+vPP\nP0+6z/6NOJKcSCXb+1OOokdp0CmlEevXr5fPFyeTtCsuj56LiDhbo0eP1owZM3J0XGJic8EpMsmX\nIgSVZbf75HD45fVGqUOHHtq/f78kacGCBbl07OWCSoLnZLf/VxERsdqwYUOBt6c065Tt27erUaNT\n5PFUFkTl8tyHFAg0VkxMDXm9sQKfbLa7BM/J76+md999L6eMVatW6YMPPlDXrj3lcnUVnCFoIXhF\nNlsfNWzYRunp6cdVr8zMTH399df65JNPtHv37pNu55NPPimX65ZcsnZALpfvpMstKBxNTvJzmA/M\nsf/D/V0s2LZtG23bdmX//lggk9jYVBYt+oaKFSsC8NVXX/Hkk6+TkfErGRlV2bLlKS666Cp+/PE7\nXn31VesgQjaDQU1CoW9JSwsjPb0jdvsHhEKXAzaczo+Ij3fSs+dBbr31O2rXrl18jc6Fa6+9htjY\naKZM+ZiKFSMYMWIBcXFxOJ1ODq36wWZLta4VHXJzL5ZEHsb27dszb97nvPzyBCQxZMj/rAN7BpK4\n7rqbeO+9TzDh+LMxnPO7ge+BDkAcweBC3nhjPPfdd29xNKPUo6TLSTlKDsplpeBh5oUsTPK5VGAR\n5rDynyQnL+SBB37G5WpMKHQd7777Br//vh64G8MhPxHoQyj0M3AaweBcli59jOHD7+H115/n1FNP\nzaVjK9Ky5QMsXvwT4eF+br31O2rVqlUobSqtcnLllUP49deuZGW9C3TBjIsbCJGaeoCDBy8GxgDf\nY7f34owzOnHbbeNzyAVGjx7Dgw8+hcvVhoyMxQSDSRjWiwhgAU7nSm699fbj8iinpaXRteu5rF79\nF3Z7LE7nYObP/+KkuKnr1q2L2/02mZnJmEQpn1C9et0TLq+kYTgmreMoDC/Mj8CtRfDeI1r+0DLd\nQAAAIABJREFUl156jZzOu3JWXW739Ro6dHjO54dbuTidXknZ8b3vWtdvsOJ9su9bKJerksLC6is8\nvLkSE5tpx44dhbyOKTiMH/+m5Wl+QzbbIwoLi9avv/5aZO8fM2aMmjZtqpEjR+r+++9Xs2bNNHbs\n2BK/os+NOXPmKBCoKzggE9vuE7SxvCKRgu+smLtqGjHizgJ9978FR5ITqeR7f8pRtCgLOqUkIhQK\n6eyzL5LP11MwXBAmu7253O6KcrtrCtKsOfETxcYmyu+PE4yxvMm5d9l6Cd4XzFXjxh2KrT2lWacY\nm+Q3K/74fME5gknyei+1djAP5PS3231jTrskc4jR660s2GLds17mUOCmnGd8vmv10ksvHVedxo59\nWj7feTKHDCWb7UW1bXvGSbUzFAppwIDB8vurKTKyvSpUqKqlS5eeVJkFiYKQk9aYWOWbMGl5igJH\nbFDr1qfLnPo8dCCrR48+OZ9Pnz5dgUArQar1+fuqUaORJMnnqyjYZl0fI0OWboTBbh+rTp3O0uLF\ni7VgwYISdyozP3j//Q/Uu3d/XXnlIK1Zs6bI379kyRI988wzevbZZ7Vs2TJJJV9R5cbEiRMVFnZF\nLtlqIqgpSJBhGZHgAdlsFYvlsGRZweHkRCpdslKOokFp1yklFenp6XrggUfUs+elGjr0Zv3vf//T\nPffcI7f7xlz6L1lOp0f16rUUtLcMt+zDebsE8YJlcjrv0YUX9i/W9pRWndKmTTfZbC/m6u96atz4\nVN11132Ki6sr+ND6LEVhYS00Y8aMnGfnzZunyMhT8yxeHI4EuVw9LeP7ffn9UcftMLvhhlssR1F2\nuWtUpUrdk26rJK1evVrz5s0rMSwc2SgIOXFgqOFqYjIv1DjZAvOBIzbo5pvvkNfbR5AuOCif70w9\n+OBjOZ8Hg0H16XOlAoHaiozsrvDwGC1YsECS1LHjWbLb/2ut3jYJKsrrra+IiNMUFVWjSD2wZRFZ\nWVnasmWLNm3apM2bN2vz5s0lXlHlxsqVK60T3r9YCuI8QQ+ZDI9NZZhQAse9Qi9HXhxOTqSSP6mV\no+hR2nVKacLcuXPl99cQ/CEIyW5/XM2bd9TWrVvVrFk7GQagyoLuMrSZVRQR0UHx8fW0ZcuWYq17\nadUpa9asUaVK8YqI6KqwsIY67bSeOfR23333ncLDYxQZ2V2BQC1dcsnVCoVCOc/u3LnTYmb6zpqv\nZik8PEYXXdRfUVEJatDgFM2dO/e46zRp0iQFAi0EuwVBuVzD1KvX5Sfd1pKMk5WTGzHH/tdgDvFl\n/xQ2jtiglJQU9ehxgdzuCLlcYbroon55eBMl4+ZftGiR/ve//+UJn9iyZYvq128lj6eynE6f7r57\npB577DHdeOONJyRQ5TiE5557TpUrV1bDhg3VpEmTnB9KuKL6O8aNGy+PJ0w+X4yio2soMbGZ7Ha/\n7HaPWrY8tUAONvybcSQ5kUr+pFaOokVZ0SmlCY8/PlZud0Beb7QSEhrnOXx31133y+n0yuUKV0JC\nA40cOVI33nijZs2aVYw1Lv06Zc+ePZo1a5a+++47BYPBPJ/99NNPuuWWW3T//fdr165d/3j2s88+\nUyBQST5frCIjYzV37lwFg8F/lHM8CIVCGjZshFwuvzyeSmrZspP++uuvEy6vNOBocmLLhyCsJztH\ndNHCqvuR8ddff2G326lUqdJR7/vmm2945533OXjwAGBj+/bt/PDDQjIyDhIZWY2MjApkZdUlK+tL\nOnU6lccff+CotHPZ+P3333nmmRfZvz+Zyy+/kO7dux9P+3Kwf/9+xox5hs2bt9GjR2f69bsCm+3o\nQ/Pwww8zceJ0IiP9vPLKc3kOpBUXEhMTWbRoEZUrV85z3WpLfmTtRHBMOTkRpKWl8cMPP9Cjxzlk\nZQlzsC8Zuz1A27ZtufXWwVxyySUF/t5/A44kJ1A6ZaUchYeypFNKE1JSUti/fz8HDx7khRdeZcuW\nP1mwYCnbtq0jKqo6N9xwNZMnf8Kvv67DZovD4djFiBFDycpKY8aML5AyCIXA6/Vx//23c+mllxZq\nfcuCTpHEm29O5JtvFlCnTg26d+/G0KG3sHLlr0AVHI4I7PYNXH/9VfTocTozZnxOWJiPm24aQnx8\nPLt27SIqKorOnbuzePFCAC644CKmT590wof6Dxw4QGpqKjExMce0SUo7TlZOvqHosvHlhnbt2qX3\n3ntP06ZN04EDB457hZCZmam+fS+zCNDHCG4TuK0to5+s0I1BMjQqYTKk6aPl90fr66+/PmrZv//+\nuypUqCqHY4Rg7D/oWvKLgwcPKjGxmdzuAYIX5Pc31V133X/UZ6699joZGplnBDfLZgto+fLlx/3u\ngkbXrl3/4dmXSseK/nCIiqou8FtxeS8JGshkmnpeHk/dPOE+5cg/jiQnUumVlXIUDsqaTilNmDVr\nlhyOSoJaggqWDswSfGbpxasFt1hzZyVrXg23rkfKJGh6SODX66+/Xqh1LS06JSMjQ3fccZ8aNWqv\nrl3Pz5NRd9iw2+T3txa8IJfrYqtf28jQ7FW0xuF52Wz9rX5+VHb7XYqIiM2hrm3Vqp2gmWCnFTbR\nVkOG3HTiHfsvwtHk5GjW823W70ZAA2AmkJEtHMDYgpCyo0CVK8eTnt4aSKVixc0sW/YtaWlpvPba\n6yQnp3DJJRdx6qmnHrGAYcNu46WX3kGaCJwFjAaeAa7IVf3dmPDr6sAWDC1LJzp0CPLdd5/nKS8Y\nDPL666+zatUvbN68js8+SyAUGoSJTEmnVq072bBhxXE1ctq0aQwc+CrJyV9ihmM7Tmct0tKScTgc\nh33G4ahEKDQTQ1UGcC09emzhiy9mHde7CwpjxowBYM2aNaxdu5bzzjsvh4rGJIi5DUrBiv5vhWK3\nuzE5eaIxtEjTgE+sOzbh9TYnJWVfmV9pFxSOJSfDhw8vNd6fchQuyqJOKU3Ytm0bCQmNyMgYCVTF\nRGDuzHVHe+APoD4mCVMPIBz4DdgM3AfcYN37MtHRY9m587cCr2dp0ykDB97A5Mm/kZp6PzbbT4SF\njWTVqsX4/X5iY2tjyooAHsbYKpvJ3s2EdUC8VVJ3oCdwLXb7o9xyC4wZ8zg2W2XgFeBi677PCAQG\nkZy8hWAwyO+//269K/aIdQyFQrz99tssXvwjDRsmMmjQIBwOB7///jtut5u4uLjj66FSgqPJydH8\n8eEYg/h3zDfCbf0UGfbuHUoodBcA6ek3cvvt/+Wjjz7lwIE+BIMxvPLKBUybNp5zzz33sM9PnPgW\nUhSQvR0zETgFmA5sBO7CCKALSLLumwA8yIYNO/KUJYneva/g6693kJJyLk7nt4RC3wPjMDlYQhw4\nEH7cbUxLS7Pemz0+FZBCBIPBIxrKUjBXmwCiSU1dhyReffV13nxzOmFhfh5++M6jLiQKCklJSdhs\nNmrUqEH16tXJyMggIyPj2A+WYNx77wOYRJRVMdFHbwFVct1RmaysdCSVG8r5RFmUk3IUDsplpfCR\nnJzM7bffx+LFK2nUqA5jxz5CVFQUAJ9//jmhUDdMFrh9GK7lLRhDLdn6uzbwHeADvrI+CwAp5J2f\nKpOZGSyUNpQ2OXnnnbfIyNgIRCN1JjNzCTNnzuSbbxYidQfGY4zj8zEc120wx8Nk/Z+NcOAx4AFC\noR4cPFiNPXv2YDLy3Qd8CzwELCMjI5WdO3fSteu5bN68jWDwIImJdfF6w0hMrMHYsQ8THx+fU/J/\n/jOMqVOXcvBgX3y+GUyd+inJycn8/PNvhEIZnH32WUyf/laR52gox+Eh+FqH6EneVu3aTWS3nyeo\nK4gWnKn69dvkcZ+npqZqy5YtysrKUkRErOAaQUPBAkGMtX3+ieA1a0vdLXjTesc7MrnWZ6tatcZ5\nyl29erX8/ngd4pc8aG0vrZBh0BghqKDNmzfnnL5NTU3V7NmzVbNmY4WFRenss/to9+7dCgaD2rp1\nq5KTk7V161arnq8IlsrrvUQ9e/Y96hZBy5YdZbIkLRRMFYRp8uTJGjPmWfn9jQQfC8bJ74/Kk/O+\nuEAJ2vrKDzIyMgROQQ1BfWvbMdbabhwnWCKns6f69LmywN/9b0dpk5VyFA/K5eTkEAqF1KJFB9ls\nNS39VkPVq9dTWlqaQqGQXnjhBXm9Z+eaf++3QgFaWHNvZRmO+dnWvHq1DGXr9YIqgqqCr6yfqrru\nukH6888/T+qA2YmiJMmK319RJhui6Ve//xK99tprqlSpep7rMNKyT2Ks/q4rOFcmc3Ci1fenCeYL\nIvT888+rbt3mgmutcbjCGgOvoqPjdc45fWW39xTcI3jDCs8YLofjPsXG1tK+ffskSdu2bZPHU0GH\nuJsz5HRWl9N5gUzGwBT5/d31xBNjCmOoihWFLCeFBvl851sG6R75/e3VrFkb6ws6V4Zgu498vio5\nDZ0wYaI8nnD5fLEKBKJkt3tlYnkirN8VBItyCeM9ljC+o+zkJYas+0n16HFhnk5cuHChIiKa53o2\nJBMztNL6/xdBhLp2PUcxMTXl88XK7Q7I7a4oE9O1TW73YLVrd7pq124qrzdaLpdfjzzyhFauXKkO\nHc5SQkIzDRw4VAcPHjzqgKampqpNm85yOCrJ7Y7VE088IUlWCuuLrC9SfUGfIk1hfSRQghRVfrB6\n9WqBzZK1uwXbBVMs5eQWRCgurq6SkpIK/N3/dpQ2WSlH8aBcTk4Ov/zyizUnPmLpt3GCME2dOlUt\nW3ay5i2/NZ/ckstYu1/Q0ZpTYyydGKPcaZehtmVUV5DDUVktWrQ5IotGUaAkycqoUY8oEGgieENO\n562KjU3Q7t27lZjYQjAzVx+eL/AIqsk4aToKshOQTLLG7D5BQ9ntZ8tud8nw/YesMoIy55i8atu2\ni7zeaJl455HW706CoTK8y510+ulnafv27Vq3bp38/uq5ypHs9qaC53PZPuN14YVlz0lUyHJSaNCF\nF/aTw+GW0+nR9dffrKuvvlrGc5s9YFvkdkcqIyNDV1wx0PryNhCMkiFC/8Ma8IctocnOrJb9/B0y\nHufq1v/fCCIUFhatVatW5enElJQUVamSaJW11hLSSoJk69mnBbFyOiMtQc5eFV6S633pAofs9get\nem2R31/rmAcH/45Vq1apYsU4RUR0UyBQX92791JmZqb8/srWF+s9qz4+9e7d59gFFjIoQYrqWJg4\n8W35fFGWQezPozCgi3U9ILu9ih5//KkCfXc5SpeslKP4UC4nJ4fZs7M9wbmz7DVWo0at5XINkzm0\n188ykFtaRtvuXPNYBUFbwVKZnbdM67MsGY+yT05nU/n9rWWzRQp+tIyuJ9SsWdFm8CtJshIKhTRh\nwkRddNFVGjZsuLZt2yYp++BkhOAyy3aJkOHuz16gnGpdPyPXeIV0yAk4VsZBlr1gyRDECR5VZGSc\nRWhw0PosyRq/xtb/nWS3X6D4+Hrat2+f6tVraWU+Xiu7fazc7oqy27PtrqC83kt0772jCmOoihWF\nLCeFBklSWlpazmnW5557Ti7XhbkEZY7i4urpppvukNt9muBnmXCNcMFNue47ILPyjZfZjnhb8IRl\nPH8v8Cg8vIs8ngoaOnRoDlH53/HCCy/Ibq8hs2I+T9DVKq+5zEovILs9LNd7p8iESGQbWz8LXDKr\n+EsFI+R0DsnxCOcXzZt3Ejwns+2yRD5fV40bN052eyUdyh4nwd1q3LjlyUlPAYASpKiOhv3798vr\njRSsFnSwxqqrDNPFVkEdmYXYG4LGatCgzbELLcdxobTISjmKF+VycnKYMWOGNSfuseaKFEGM5WxZ\nIxO+FyuTZMkj42TK7TSoIZhoGWY9BH2s+e5Cay68RvCDTMjjdcr2XsJBORzuIm1raZCVDz/8UF5v\nPRl2i5EyqaifktnVTLL6rp9Mhth06/8dMs4cn/U7XMYxN0VmJ+BsQVAOh08eT/1cYyeZ3XCPNTYN\nBMkKDz9dM2bM0KJFixQXV19ud6zq1GmqmTNnKjq6ugKBpgoLa6zWrbscc9f7cNi1a5fmzp1bYpO6\nHU1O8hONHQNcByTkul/AwJOVsPzA4/EAhnGiX79+PP30q2zb1peMjDp4vW/y/PMvMWzY3WRkvI8h\n5/BiWAomAX8CL2JO5UZism+3w2Tjbge8i9P5FF27nkXLlg2YPn0fn332LSkpaSxYsIqUlBSuuOIi\nHn74fhwOB9HR0Xi9DUlJyWbDOIA5tHALcCrQG7s9g1BoKSbr92nYbNfj8ZxDRkYrvN63ycoKIyNj\nJjAUmEMw+D5VqjyLJDIzM3NO7B4N69atBZ4A4oAdpKZWYu3aXzHDYs91p53DjX1mZiYffPABO3fu\npHPnzrRo0SK/w3FU7Ny5k3HjxrFp0yaysrIASuxBt3nz5nHjjf9lz5699Op1NmPHPsq2bdtwOqOA\nrcBqTP+uA3YBdYFmmFPdpwEJ/PXX9cVV/VKNI8nJ+PHji7lm5ShpKE06pTTB5B5wAp2BCzCH8UJE\nRUXx+++zgdcw6RMuBd4BFgB3YObOWZi57yvgSuAj4GzgesxhMo91/yLrbSnWzydAFJUq5T4UXTAo\n7Tplw4YNBIOnAPOBUdbV2zCHyO8B5mBIx/ZibJdzgPcx/ApTgDOt/wcAc4GrrXIWmz2A9D+AMZjx\nfA/IwtgGIeudbiQb69ato2/fq8jKugFoyB9/PEqfPv2w2aJJS9uC3+/j0ksH4vP5jqt9c+bM4fzz\nL8HhqEtGxjqGDr2OJ598+Pg7qpiQH42zAJgHLMX0Kpgefr+wKpX9DkmEQiGGDh3O66+/giT69r2M\n9u1bceBAEmeffRannHIKdeu2Zt26h4EuQHOMkPQG3gDeBg4CQeAm4CngM2AYTuce+vbtQ+/eZzJw\n4B2kpLyJYTi4DaMc4vD7b2bo0K488cRDrF27loYN22AM7jDr3nRgE3A7sIzq1bexe/dfOJ3tyMpa\nyYABfWnRogHbt++gbdtTOPfc3oRCu8gmFbHZWnPHHWfy4ouvk5Kyn4YNW/PZZ1OpUePIWcJttorA\nA1Z70oCOtGoFa9duISUlAngc2A7cTe/e3fnww0NDlZmZyWmn9WTVqjSysppis73PhAnPc+mlJ588\no3379nTp0oXWrVtjt9ututro27cvlBB6HoCff/6ZNm26kJLyElAPn+9uLrmkJi+/PJbo6HgOHszC\nLDKaA19glNFTmFPEnTCiP4V27Sbxww9fFmhj/g04kpz06dOnRFE5laP4UVp0SklFKBTilVdeY968\nxdSrV5M77hhOWFgYmzZtolatZhgdVxPjVFqEz+fFZguSkuK1rjkwxm88pruTrWeeBC7DzIN+4GfM\nXNQQY8z1wDBIgZmPd2HoWe/Gbv+OZcvm0rx58wJrZ2nVKdnjM336p8yb9x3BIBiisTBMf1bB9P3r\nmMXGpcB+jD3jwTAzbcxVYi0M1e0G6+9l1r2nAysxYxMNpOF2Z2KzNSY9PRNYAtjxej2kpfXGGOhg\nxrUjhrJuAuDF6x3ImDG3cMMN+XMUSaJSpTj27XsLIxe7CQTa8MUX79ChQ4djPV5kOFk5OT5i4IKD\nJOmpp56R399BJj5qv3y+7rr33gfzuMwfeughmVidAToUd5MdwxMrmCMTVxxrhStIDscgDRt2mySp\nd+/+gtd1KG75gVxlrFRcXH1J0vjx42Vie66QYaloZG17RMnEa43QqaeeoY0bN+rDDz/U0qVL89Tz\nr7/+kolxTctVfns5nX7BEkFQdvvDatSo7TG2CAKCTbnKeFDh4RVVtWoDmdjp86wtmOt088235Xl2\n6tSpCgvrqEOxTEsUGRl7vLsUh0Xz5s2PUN+StfX1xBNPyOXKHZqzVYFAZUnSpZdeZm1xRQkey3XP\nb1a/dxPUlcdTUd9+++1J9de/FUeSE6nkyUo5ihelRaeUVFxzzRD5/e0Fr8rjuUJNmrRTenq61q9f\nL5crRjBcJlb5dMFfgq/ldofLhCnmPhiWKMMEFGHpxy4yB8bPkjkX86tM+EYfa579KJfu/FhwjvX3\nXoFXXbqcW6DtLK06Jff4OByXyByCbCATntneGptZufryNRnmkSyZEMFIGbIAc2bLzFF1BWEWmUGs\nTGKy7OeHC24W7NKhZFpXyoRzzJAhMxiU6/6NMiEd43Nd+0xt2pyR04ZQKHTUNiYlJcnp9Cp36E5Y\n2BV68803T6rvChpHkxP7kT7IhZnA4YmKiwCffz6PlJSbMRyCEaSm3sbnn8/N+XzJkiU8+OAYzLbQ\nLiCbSxDMCiwLs7LyY5KO9ANOITb2a0aNuhuA8HA/Nls2b7IfyM2hvAOv12wzrF27FrMFPwmzzTTX\nKv8v65m3SEpKonfv/vTtewk9e/blq6++ytMem82DSXjyDfAg8DN2ewtMqIadUOhu1q5dTnp6+lF6\nxYnhhAazBfYeEMTvd2GSYgzDkJG/RyDgyfPkrl27yMpqzKGhb0JS0m5CoRAni/POO49PP/30pMsp\nbPj9fpzO3OT5O/B4zBgbLsosjMd/MoZfWxh+y3DgP0Aap5/egY4dOxZpvcsKSouclKP4US4rJ44D\nBw4wadJEK1RwEOnpk9i8OcjcuXOJiooiKysN4720YcIoKgPdcDgaYPTeEMw8NRQTuggmxLAhZqP5\nO6AGJjSgLoZP+S6M/pxg/c7+u5n1/A7Ax44duwq0raVRTv4+PsHgZEzys0TgWUx4RXPyJnrZjrEV\nHJhccJcAl2M8zU2BepjQQQ+hkMMqq1mu55th7JUozNgFMLvTr2HG+UyMffMiMBu4EOOBzjtf+v0+\nVq1aRZ06LXA63dSq1YRly5Ydtp2BQMAKt5luXfmDUGgOTZo0Oa7+KulIxoRcpGG+PUkY66ywIUka\nOPAGOZ135KxE7PaH1avX5frkk080bdo0DR48TCb4PXvle5q12n1a0Npa8fa3PIQxggry+yto0KBB\nGj16tDIzM7VixQo5HOGCuwS3WyutmwSPy26vrKlTp0mS3nzzTeU9dZomc+DrLmslfaY8nijZ7Y8I\nUgVfKBCIyjkcGAqFVLt2U5kDYp0EPeR2R8rvb6hDAfrL5fdXPOoqzXixw3WI2zdcderUVWRkvIxH\nuZsM5+JQ9es3MM+zK1eulM8XLXPIIkVO53Cdemr341p5rV27Vu+++67mz5+fp56BQEA2m00ej0dh\nYWEKCwtTeHh4iVvR7969W1WrJsrlGiR4Sn5/Tb300quSpKFDb7JkpZ/MgZVsLstwa7X9uKCpLrjg\n6FzX5TgyjiQnUsn2/pSj6FFadEpJxK5du+R2R+oQI4UUEXGGZs6cqTVr1sjhqGZ5Dz06xOGbJK83\n3trlrCjDCNVZxmvcRYYq7qA176VZ8801uebEF2U8ohEyu6/Zaa37yjAz1BbU16233lWgbS2NOsWM\nT0Se8TE7072sOT7R6vfsdOB3WPP9RB1iGGkp43n+j4zH/00Z1pFLZA4C3mPZLH8JfpfZcX9F5kB6\nDctGmijjyc7epf5WUEk2W1XVrNlA4eEx1nvvEjwovz9KX3zxhSpXri6YYNk676pixTjt37//sG1d\nsmSJKlWqpvDwevJ4IvTEE0+fcL8VFgpZTgoNkqStW7cqNraWAoHzFQhcpIoV45SY2FRhYR0UHn6e\nPJ4I5WW4+MIa1FYyWwwBGWqVFBnX/8XW52cJ6qlSpZq6995R8nhOEwyxjKTmgnpyuyvo7bffzunI\nn376yXp2jMyJ3ossAftAJrwjXHZ7QLm3GFyuszR16tScMjZt2qRmzTrIbncqKqqmvvjiC/Xs2Vdh\nYS0UCAyQ3x+jyZOnHHVAnc5K1hfEYX2h2qlv377y+arq0KIhVdBeZ5/d8x/PT5/+vipWjJPD4VKH\nDmdqx44d+Ram996bIr8/WuHhFysQqKOrrx5yzK0XSqCi2rVrl+69d6QGDbpRn376ac71pUuXymYL\nsxS8x1I62dzUF1hK36PWrYuW4ujfgpIoK+UoeSiXk2MjFAqpY8cz5fEMEPwgu320oqNraO/evfr1\n119lFv4LBS/JsDddLKguuz1Sdnui4ExrfqlozXkrrPn0bcvI6qNDiUg6y3D/+izD7nyZ3AJrBb1l\nGIPClE27mZ6eXmT9UFJl5dD4XGWNz2OW/VBdsFOHFh6RMguPqlafR8mESzSWoY873bJLTrHmJ7+M\nsd3MKutCay7zWJ/ZBPUEqwTvW9ejc9lQEpyiwYOH6IsvvpDfX1lebwc5nTGqXr2+Fi1apOXLlys8\nvFGeZyIj22jBggVKT0/Xxo0b/8GMkZKSotWrV2vXrl3H7JvMzExt2rRJBw4cOOH+PV4UhJxUxByB\n7ZLrp7CR04C9e/dq0qRJeuutt3T77XfK47kilzE6SodWOy9YwhRhfbnXyPASvpBrQBsIpuvQiuxU\nVatWS4e4j/cpOxFJ27Zdc5JKhEIhnXJKR5kVXxNLsKItITNchg5HBZmV9nod8jjX1PPPP5/Tlg0b\nNqhatUS5XBXkdHo1atQjCgaDmjlzpsaNG/cP/ubDoXfvi6x3VrIUU0ArVqxQbGwdGdLxGjKryZYa\nNuzWI5ZzLAP378jKypLXG6FsTkxIUiBQR/Pnz8+5Z8+ePVq4cKHmzp2b80MJVVRHQmJiHUvh3Cez\nMk/UIYqeHy1lH6MFCxYU+Lv/LTicnEgld1IrR/GhLOiU4sKBAwd01VWDVadOa/XocaHWr18vSfrt\nt9/kdmfnFqhpzVte63cTGR5eycTHVpCJje1hzTndrHkne/75Qoa7f5xMFtwK+meM8tnWnP29bLZA\nobS1NOqU7PFJTGylHj0u1IgRI2SzDc7Vdwet+SZRJgtvhIwDsJmMMzB7nF61bJ4oGY9x9hmt7MVL\nhA5lIn7buq+F9VmCJQfjrGdmC3yqVq2OKleOl3G+LRFkKBA4VZMnT9aQIdlOyL+UHXvu88VoypQp\nqlChivz+eHm9kRo58gF9/PHH2r59e777ZPXq1apSpbb8/mpyu8P0zDMvnFQf5xcnKycoGzkJAAAg\nAElEQVTXAaswCd+/wSR9//rk5euYOGxjrrxykMwqS4L9MgfXAjIH+a4T/E9mezzC+uI7ZFZc2dsb\n4YJtuQTxLrndfnk8l8mEboQEtwoGyOPpp0GDbsoZPLe7kqC79dwkS6HstZ4ZLputjlWXeMENgtZy\nOmvrnXfe0a5duxQKhZSQ0MBSJD0F8XI6Y/Tll18e14C6XFHWO7I9nxXUr18/DRo01GrfaYKGstki\n9dVXX520AGVj9+7d1lbRoVVkeHgfvffee5Kk1157TU2aNFFkZKS6du0qr9erbt26lWhFdTjYbG4Z\n/tArdcijnHu17Rd01vTp0wv83f8GHElOpJI9qZWj6FFWdEpJQ1ZWlmJjE6y5KDuxxWiZ0MPch7mS\nrTm0ioxhfJcOhRyGCewyDqfs+6+x5qSrrHkxZM3N1WV2bGNkszkLvD1lRad88skn8vka6ZBjZrKM\n57iajOPvEWuOryXDs5zd76tlnGPh1t/Z15+SOdwXJ3jLKkcy/NmLZHamR1qf17XG02fZMfWtue4c\na/xuk9t9kwYOHKhAoIFlJ9UVXC+7PUGDB9+iChWqCD6x3rFCEKawsNMUHh6jH374IV99kJDQWGZh\nZg4T+v1xWrx4cYH28+FwsnLyEyZKP5v9ogHw4UkKV35w2MaMHz9Bfn8rmZXMhTIe44tkcs1nyaR2\nbCy7vapstgiZVVM7SxDqyRjQt8gYxb8Lqshud6tVq85yOBIsoWwqQ+b9jRo1MlvsCxculM9Xy1IO\ng2QM3cdzCeQvgupyOmMsZfKs4Ek5neFyufzyeCqoZs1GMh7oxTp0Ariqhg0bdpwDGi5jsK9XdjbB\n2rUT1a5dd5lEJMZb7nb31JgxY09agLIRCoUUH19PZvWaHU8dnUMg3rhxY6WkpOScQP7555/Vu3fv\nUqWoJFmTgMeaGHpZY77cavPLgiqy2cK1bt26An/3vwFHkhOpdE1q5Sh8lBWdUtKwd+9eeTyR1rwl\nGedRtEyyisrW3BKS2bFtbxlPE2R2K/+w5t4mMtv9T1j3rrWezd5lrS3jCY205qyhstmuUatWXQq8\nPWVFp4RCIV199RDZ7dke38oy55kilZ06vH79JtYip5Zlp2TKLFDCFB9fX07nNda1nTLMXO9a9s9a\nywCeYY35dTKLIK8OJS65xLKLxlnjuEiHdtoT5PFE65prBspmu8e6/qXgQXm9EVq/fr0Cgewsx9k/\n3WVYOaaqdu1mx2x/WlqabDaHcoev+v1Xa9y4cQXaz4fD0eQkP6wXaRgvMphsHmuB+icvXyeGq68e\nwH/+cwYORxzwMTAOwzG4AsP1VxfYQZUqXr7/fhadOlUGfgT6YLiVQxgGAx9QG4glIaEB33wzk8su\nOw2XqwaGKD0G+B/r1v3Kb7/9RlxcHGlpOzC8zAFMIoqZmFO9AF8CPrp2bU98/Ae4XPfg8YzCbveR\nmfkj6el72bz5Usxp1TbWMxWAprhcruPsBWG4nj/G8CVez759B9i8eTNwhnWPg4yM01m3bvNxln1k\n2Gw2vvzyI6pXH4vLFY7P15UJE16kbt26AHi93hwi8rS0NBo0aMAvv/xSYO8vCgSDQapXr4lhTqkG\nzMCwk3TBiP8IYD9vvPEsiYmJxVfRUoyyICflKBqUy0rhYNu2bbhc0RiWBDB8vQ0AF2bKb4BhgHoG\nM99Vxkz9aZg59jMMf+/TmJwDERjWBWH0pA3DNlXT+tsGfIvNNo1Ro0YUeHvKipzYbDYmTHiJRYs+\np1+/5jRsWBuTwuIcAMLD3cyZ8yVDh16P3Z6GYa4IxzBKxLFly06ysj7AjF08JjnJEmA3hg1DwA0Y\n++dDjA3yJSbXxGRM8pEo4CKMeXiKVbNIoB6XX34+Xbueht8/B5MApTuQQK1a9YmNjSUYPIAJQADD\ncPITJlfd6WzbdmxbxO12U6FCLCZ4ASAJm20BtWrVOq5+LGjkx1D+AxOjPAPTox9jMmwUOSQxceLb\nbNmyg+uvv4FAoAKw2apee4xQLADeYO//2bvu6Kiqr7vf9HmTSSekklAloRdpAoaOSK8WRBGRIlWp\nIkUsCIifCiJYQFFREBF/gAWVpqj0IkgXpEgLNZA+s78/zptkAkkIJCGJzF5rFuTNu+/dcubec889\nZ5+LV2G1WlG1ahUIpc17kB+1BcBSAPshCvNhnDhxGP7+Qfj9923w9d0LoDqEZHsZUlMHoVevZ3Ds\n2DF4eUVDlO03IMKwF0K/ci8U5XmULJmMjz+ejWPH9uHs2eOYPn0KUlNbAiinteB5iKB+pv39F/T6\njejWrRuefXY0unR5HO+99wFkc5NjT0Bo7g5ChH0BvL29UKdObRiN70I2Axdhsy1E/fq18tDjN6Ji\nxYr455+9OHv2OOLjz6N7927p30VERODixYvo2LEjWrRogfbt2yMqKipf31+QcDqdaN26C86d8wPQ\nE5JQZhOA4RAKPh0AC2y2KISEhBRiTYs3iruceHDn4JGVgkFUVBR0ungA32lXNgHYDpOpL8xmI0Rh\nHgHJatsA4nn5fxB6zCsQujAfCH2YF2RufAdC0VoDYkDyhyhdDgil6Q44nR9hwIARGDFiLLp06YXZ\ns+fkCzXpf0lOli79GlOnzoKXlw9WrlyE775bCj+/X6HTbYbD4cAff2zE//3fLDidqyFj0RTAaMhY\nVIMou19AjHLLIEpwDER3MULUtxgATbR/G2pvbg9RsP+BjKkeQhUHAHthte7AqFEj8Oijj6Jx42DY\nbJXh49McPj6j8Nlnc2Gz2TBv3lyoalOoajPIZusJABWh189GlSo310UURcGXXy6AzdYD3t4toaqV\n0aNHczRt2jRPfXqnEQvpzZvnWRZEQLYGeyBbiyHadX+I0n0AkvrMN4uyN5jGJ0x4iapamcAHNBiG\n0m4vQas1jBJ5660dGYi5Xq9/jq+88gqbNGlJCTA4Q/Gt6sQMgvUzFEqVbZqpf5rmrtFGO/aIIhBG\nnS6A8+fPp8USQOAfAi9ox04+NBii2LVrVy5dupRXr17NVN+xY8dSfHgStXqtJWChr28oDQZ/Ggwq\n33lnNkuVqkijsT+BD6mqtThs2OgcjwgUxYfACro77MfGxjIuLo41ajSk2exPo9HGgQOH33LAXm6w\nZcsWzp49m8uXL8/2+WvWrOE333zD5OTk3Bx95auc3C5++OEHenlVY0aAhOv4KZBCxdeZQDCt1qc5\na9adCTD4r8NdTshcHZMWCVnx4M6jOM4pRRm//PILfX2DabWWpNXqy/Hjx3P37t1cuHAhDYZybsfn\nTgIR9PIqQb2+IiUGpjolCNC1xi12u78Cgcluf7sSggVp66+OJtNTBOZRVeuyf/9h+dqu4jynzJnz\nHlW1NIG51OnG0ccnmGFhFagoM7Vx2ESrNYB6vZfmSjGeGQnPwgnsdhszK4G9zCAuqELxUS5NccWo\npo3JGe2ewxQ3jJEUv/UgiquGNwEzhwwZxrS0ND7//CSWKVODkZFV2LFjR3744YdMSUlJb8ORI0e4\nYsUKDhw4lEajSoulBMuVq8Zjx47luh9OnjzJlStX3pC0rSCRCznJEt5uwpLVJzcIhphnAdmi7Iew\nlU+DZAgBZCv0WhZlMzXC6XTSZPIi0I/AIALrqKqdOXLkSI4bN552ezCF6kZ+nBbLI3zzzTfZtm1X\nAjPdfrRbKX47LSlRvAEUDsndFC5jPcUn9SsK1/C9BB6hxVKCTz01gOIrVF971+dUFC8uWbIkS7qb\n2bNnUxziK1AocgIJKLRYAmix9KDNVp9ly8bQy6ulW/3O0mAwMy0tLdsBVRQ7M3gvSWAiGza8P72f\nTp8+zUuXLuWrELnw3nsfUlVDaLU+RZutGjt2fCT9XefPn+f58+e5a9cuDhw4hD179uGSJUtyI4D5\nJid5wcKFC+nl1fW6BcJMCaJI0SYUC63WiExMHx7kDi6OTZecXP8hczVZFQlZ8aBgcTNZ8chJ/iAl\nJYXHjx9PX782btxIVfXTlCSXgSeBQEnqdM9o6988CnNCtKZIlaIEPu/WPoEUylTXPLpMU8w2UwxU\n3szwQb1Avd7MpKSk26r/f21OCQmpkEmP0eufpF5vc+tLUlFaanqFi+N6IyXoz4sZQXBJFCOge6Bl\nR8oGxqDlUnAF7vlTuJbt2t9jtOfEauP6nVbOxsDASBqN5TXd6EsC3rRaq7J+/eaZlGUX4uPjefLk\nSTocjnzro9OnT3Pnzp03GCbzilzISZZwpbk5CnFGuv5zO1gGcWrZB0lSDoiQ7svi3kyNOHjwoNsg\nTidQkiZTU7777rtMSkri4MFDaTKFEJhCo7EPQ0LK8vz585ww4UUajY+6/TDfoUT6uhgqXAklylGo\n48zM4CJ2RW7eQ2AKmzZtrQmju5I6lEajPyMjY3j8+PFMdX777be157+tTSy9tfKrtbIOms3RNJvb\nuz3vGvV6U5ZClzGgPgR6UCJXdxAoyYYNG+Wr0GSF1NRUmkw2ZgSAJNHLK5p169YlSUZGRjIiIoI6\nnZ4u+jpxzL9lAbxtOckLjhw5QkWxUSiRrlCsJSWYke57OQE7GzdumW/vvJvQpo1wekdGRjIqKuqG\nD3lbk1WhyIoHBYubyYpHTgoG1as30taqHhSD0DRKMFlPSkBzgLZeVmIGv2+opmyV1NZWlWJxPkYJ\n/KuuPYsEzlOCpV3rXRINBssNnLu5xX9tTilRojTdWSsUZST1eqvbtWsUq/AayobkkFtfDtP0l5cp\nrFt+BJ7T1rLVmu4xQRuf0tpYjdLue4Bygu5a705r190V7XsphsrKFEs2KVS646iqjfnpp5/ma19k\nhVdfnU6z2Zd2ewx9fUO4cePGfHv2bchJgSAK4vxiB3DR7bpy3d8uZGrEkCHPUZRk16B9Q53On/v3\n72eVKvXo5XUfLZZGNBq92a/fgHRS68uXLzMioiIVpTrlyEilZJ4hhWg9kBnJSEZowjHI7T2rKUcU\nzzEqqpw2IWxx+/5hAq9Sr5/Ipk3bZ6rznDlzaDa30ATQleTESGG7kPIGw9O0WPypKG8Q+IVWawd2\n7PjITQZU1SYrV3SxhWFh4bkWiMuXL/O116ZyyJDnuGLFilyXu3DhAk0mO90jUu32Lvziiy/S73nx\nxck0GJ5x65+1tyqAeZKTvODnn3+m7LIDKbvx+tokH0w5XjQRMLFyZU+ykYJCcZEVDwoXHjkpGIhF\nc4+mID1EsRa/Q2FRGK6tPd9rc/vf2vrjoylmTsqpbF2KUcuifbwIbNCUtH4EjNTpphH4hRZLVz7w\nQJcCbVNxkpUxYyZQVetqivDHVNVAvvLKq7Rag+jl9ShVtQL1ej+Ktb8sM1gpSNnM6DWdZBKFpcSV\nUbaUm77wtXb/CW1tq0lh1HheG98GFB5sM4WCV4x6cjI+kqJEWzQZGUBRzJ9g7969CzSRzObNm6mq\nYQROanX6ikFBkfnmXpqTnBhyEJiaNxGorBN7Zw0vAF8BGApJge2OXAnytWtJAELdrgQiLCwU//vf\ncuzfH4SUFH9IgF0l/PLLFgQGBgIA1q1bh5MnT4NMhTiy+0OczAHgMYg70qMAEiBBCwSwAIAZEjX6\nGiTycw7S0oIgBCCdIawT+yGBgS3gcETjzz8XZKpzgwYNoNdPgBjnq0OvnwSr9U8kJ09Caup0APth\nMi3H/Pnv4oMPvsCJE4vRvHkjTJs2+Sa94YAE7I2HsF58jStXru/W7PrxGmrVaozjx2OQnFwNH3ww\nBC+9dBjPPjvkpmV9fX0REVEaR468DqdzOIDf4XCsg6o+np7n/e+/jyItTY8M8Tib3eOyQp7lJC/Y\nvXs3xA3NCxLdvQciMwkAOgD4BoAJiYl3IoP7fw8uGckONWvebMrJhEKVFQ8KFjeTlVuAR05uAc2b\nx+KLL0YiNdUG4BRENxwH4CXIXKgAWAdgOoRxoQJkDbpP+84ECeAarT1RgaxXrQE8CTHQmlCr1gpc\nvboUsbH1MWPGK7dd3//anPLKKxPh5WXD55+Pg4+PN6ZP/wYNGjRA+/btsHXrVoSFPYE33piD1avv\nQ3KyA0BbyPgchvAsfAJgMDKYLBRIgN4lyNgkQYL6XE1KhOSSC4aQFPQAcAIyhgYALQD0ArBWuz8F\nGXIQDWG2CAFwBvPnE8uX/4w1a1aicuXK+d43e/bsgU4Xiww9sDPOn38ES5Ysgc1mQ2xsLFRVzff3\nAjkrym9AesYKoBaAXdr1qhC+kfq5fIcRInyfQI40AOndYACnIb2cpTY1adKk9P9XrlwOqjoVCQn3\nAPCHqg5G//698Pffx5CS8gfkB/sygI3Yu3c64uLiEBgYiL59h8LpjIaLvg0YABGkhZAoXhOEeqUk\ngGehKGkgLRCWjBCIEJUGcBBJSZcgwnEYMhGYIHEAiwFsQkRETKb6V6lSBR999A769GmDa9cuoFq1\n+/DRR9+jf//n8McfKiwWO9555010794d3bt3z2V3AqLEvwyZyCoBuAKrdT0AYPv27fjmm//By8uG\nxx9/HCVKlMhU8uuvv8aJE0FITv4MgIKEhC544YV7MXz4YCiKkuNbhR5uGdq2fQh7946Fr29JfPbZ\nAkydOhWKoiAxMRFbtmyBohDkEsiPM9fUd/kiJ7GxsYiNjc3tOzPhnnvugU4HOJ1nIT+NtyCR3cMh\n8vI7gNqIial3W8+/2/Hss8+my8nWrVtRtWpVxMfH4/DhwwgNDUWfPn1y+6hClxUPChbXy0rp0qVx\n9epVnDlzBqGhoTd/gMAjJzmAJL766ivs2LETFSqUx6OPPorx40fg889rQ5SvKgBegCjC90JoyhSI\nAWEkxIX3EMSmNhgy3z8NoZRrhIx0C520797U/o4G8AH++mtdntvwX5tTdDodxo0bhXHjRmW6Xrly\n5XTlkyR++ulhyPq0G8CLEP3lKQAPQ7xlvwXQDcJEMg1AHIAhUFUbEhJWAWgJ4HWIsXCm9pYaEMPg\nrxA2jDMQitQ/IWrfCcheojpEZ2oHMTQOhLB/TURcXATatu2Bo0f33HYfnDlzBh9//DESE5PQuXMn\nVKlSBYCsz07nRK0tgQAWwuk0ok+fdwGkIjBwNDZvXouAgIBcvWft2rVYu3btbdfzeiyF/GJcqAwR\nqNxAgZhn/++669OQseUcg1w6yS9btoyVKtVnmTI1+PLLU+lwONirVy/tmNx1REACLRkdXZNvvTWL\nqhrCzCmst1F8c+7VjoTc/ZE3UFH8tKMkHwK73L57k2azn3Z00Vk72mjDDP/VOaxcuX62Zv3rg/PS\n0tJu+8hAjj1KaEcljxIoyapVa/CHH36gqpagTjeGJtMTDAqK4pkzZzKVnTRpknak5mrXVQKGW3K2\nP3HiBJcvX85du3Zlut6pUyfu2rWLq1atYtWqDRkWdg/vuSc6NzvxfJWT28Xp06ep06naEZZ71qMf\nKC43KwjoOWbMmHx7590Il5y48Oeff7Jz584kc3VMWiRkxYM7g+xkxSMneceAAcNps1UlMIFGY20G\nBJTmfffF0mLp5Tb3/asd35PCoGCmBIr9TXFR+5cZbhhWiv9sIG9MYR3j9vc6BgSUy9e23E1zSseO\nPZmRplpcECS5SwjFL7wXgQ8I1CKw3u2+lxkUFEGdzoviTmqj+KC7vt9MYcYgJeX4u9pYRlP80O3a\nx4fiivMaAXc3y30U5g0dW7fuzFmzZnHDhg05EhNcj5MnTzIwMIJGYx/qdKOoqoFcu3Zt+vdjxkyk\nxRJIb+86NBj8qNe73HGdNBqfYf/+Q2+7X3MhJznir1xeywoNIT4COwBs1z6tIf4PP+EWaVd27drF\nadOm8d13302PdpWc4yZKkIBrwJoQGEJVrcXw8DIU32CXU/qLFJ+cdZT0jePcyq2lKNFztH+fowQO\nvk2gCX18Aik+PL01xSlzZr6goDK3PUi3Ap3Ol5KP3fXu7mzWrBkrVqyTaYIyGPpzwoRJmcpOnjxZ\nE/ZF2sT3EAFbrhXlFStWUFUD6ePTgqoayuHDx6Z/Fx0dfcP90dG5UpTzVU5uFzt27KDFUoHib/ea\nW/+u1OQhiEAJjhzpUZTzguzkhMzVZFUkZMWDO4PiPqcUVZw6dYpmsy8z4mUSCURQp6tJRensNvcd\n0+Y+UrKzmSkBZeuZEaDn+rj7wfZlRgrr7tr1YVq5SvT3j8zX9txNc0qXLr0oFLXTKUbAD+nvX5a+\nvmF0xdEA7QjUJvCz2/iMp8HgxUqV6lOMfPdRfJR/IrCTYjx0xeT4ElhCUbhbUIL3wihsGL4UX/Sp\nzJzufDclxstGMULWoKrew8aNH2BiYiIvXrx4U+PgiBFjqNcPc3vmItaocX+me/7++29u2LCB1ao1\nphixXPcuZrNmnW67X3OSk5xcL1zYBUl99ylk5/UIJNVdbvArsk9q0jyXzwAA/Pjjj+jQ4RGkpT0K\no/E4pk6diR07fkPjxvXw/vuLkZbWDuJatFGr3lIkJFxGQkIlAFchLgreEB0/CXL0EALZRPpCTljG\naN99ozX1XcjRxCkAm9C7d1+8+eb7kKOnwQA+hvhd+cNofBt169a5lSbdNux2Gy5fjnK7UgH33uvE\nvn1LIbEIgrS00jh//nSmsuXKlYPZXB7JyW/CRQ6vqibodDfPPeNwONCjRy8kJKyAeN5cxNy5NdCj\nRwfUrVsXVatWxVNPPYWePXuCJBYuXIhq1aph7969N3t0vslJXlC2bFkYjReRlBQNOc6yQlwvRkD8\n108AuAQ/P+8cnuLBzZCdnOQSRUJWPLgzKO5zSlHFlStXYDD4IjnZpftZAETA6RwB4HHodC/A6awC\nYBKANto9uyAeCm0hasBuAOshWUu/g7hXhEHW1TUQj82rkIxwT0BcFh8AUAv+/rlNxZA73E1zSvPm\n9fHVV88BeByyhn+Pt9+eg8uX4zF48AQ4nYmQcbFA4q+mafe9hbQ0O4KCfLFnzxoAz0LGsSvElSEe\n4kazGBJ/1RuSkfECpItWQ/yg/SBuFx0hiU1CIa6v4yC6VCjE13kwEhLSsHHjA/D1LQGnE/DzC8R3\n332Vre/4+fNX4HBUcLtSGpcuXc50T+nSpVG6dGk0a3YfDhx4F4mJ9wNIg6q+h/vvb4bCghXSo19r\nn+GQEShoZNL2y5atzowkG6SidOXgwYPpdDo5fPgYKopZ2+kEMINK5bi2u91M4DcKg8VLBAZTIjfL\nUlFM2u7Hh3I89BvFtSGAwk0o79PpnmSbNu0pFuX6FItyRwJmKorKWrUaMy4u7rZ3M7eC3r0H0mJp\nS+AIgbW0WoO5YcMGDh48klZrCwpx+Aaqahh/+umnTGVPnTpFvd6XkoN9NIEAtmrVNlfvPX/+PE0m\n70xWBLu9Kz///HOSZEJCAmfMmMGOHTuyY8eOfOONN5iYmJibHX2+yUlesWDBAsoRopViefem8IZS\n20VHcNq0afn6zrsN2ckJWeAUPYXccg9uFf+FOaUoIjU1lZGRMdTrX6KwH7xLOTa/RJutJps3b8cW\nLbpw8OBhNJl8qNN5U1GsFLfDKZTEIa2YwWpho/D6lqEc30cT+IhCufql25rxGAED/f0jGR4ezYcf\n7pN+OpwX3E1zSo0ajQksdOvTPmzfvjNjY9trFmAfAge1sdBrVt5HKW6nFzQdxk/Tl1ZTXDXu0cbw\njNtz21JRbLRYamtrYWmKW+I87dl+FEpAH4o12kbgT+27/W7PmUqhGxQLcUBAeLa82StXrqSqliLw\nB4GDVNX7OWrU+CzvTUxMZJs2XWk02mg0quze/XGmpqbedr/mh5yoyKCKuFPI1IjAwEht8F2dP4kG\ng42HDx8mSV67do1//vknvb1LUqebSmA5VbUedTp/io+Uq9xgyrGFDGC1anXZqlVXbaJwEmirCUAZ\nCm+kq9wM6vV2Age0v/fQxT/YsGHDAsmAlx0SExN5//2tqNd702IJ4qxZ75Akk5OT+dRTg+jtHcKg\noDL8+OMFN5R9//33KUcopSgbA2/qdOZc1d/pdDIoKJIZG4gDVNWS3LNnT/o9165d4969ezOVQxGf\nqC5fvsxx4yby4Yf78NFHe2oTyf0EjlI2RE43OWjCmjUbcPPmzTxx4kSe3323Iis5IYu+rHhw51Ec\n55TigH/++Ye1azehXm+nooQReJdG43BGRcVk4jVOS0vj6dOn2bBhQwLuCZl+1ZQnbwq1qpPi4tiZ\n4uMaxBtjfaZrCtVsArtoNj/O2NgH86U9d8ucEhlZhZI4zdWn/0e93p91695P4A2KAa8TheZWZWb/\ncFJo5T6nuKAGa3pMIMU46Hqug7IBWqP9fUb73pUMJVErd1K75qv9/T2BLm7ycJHCuf1p+vtttige\nPHgw2/Z98ME8BgeXo59fOAcPHnFT5ffixYt53mzFx8fnWU7aQ+zwR7W/a0Ds7wWNTA3p2bMvxa/m\nDIXHOIxG4wOcPXt2pvv279/Pjh0fZf36rTl16gx26NBJ+7FOIPAUJQjuJAEnDYZ2nD79dQ4a9ByN\nxgHaQEZQAvWe1d53isBOms2laLWWv07ghFd55MhReRqkW8WIEaO0dswiMJw6nRePHj3K48ePMzIy\nmnZ7NFU1lB07PnKDI32XLl20NgZQHPpLEsidokySW7duZWBgBFU1nGazne+992H6d9988w0rVKjA\nyMhIkuS2bdvYrl27Ij1RJSQksEKFGjSbe1F802O0CcKVMrS6Jg8XKNmmrFQUld7e1Wmx+PHVV6fn\n6f13I7KTE/K/t6h5kDcUxzmluOB//1tOVQ2g3V6Ver2dpUvX4GOPPX1DALgLI0aM0JTfXpRYn5IU\nn9Vgykmsa138kEBVTXmqQ8n6doqSHCuSYuU8q92bQoPBkucsa3fTnDJo0AgqShNKIOUuymblGTZr\n1oE2WyAVZTjF/ziAwJOUjcyH2hr2tnZ/EoWDuSp1unbU6VpReJfDKVn42lNOVd31nQeYYclO08b/\nH+1vi1Y2SFOSw+lKfS3W6DjtvsM0Gr0KLHPw7WDz5s309Q3Js5xsgziebHe7tm+m+X0AACAASURB\nVDsvD8wlMjXm2rVr1OtdbBQRBD6izdaO8+fPz7ETrly5wvDw8tTpwqgoEdqAtiBQhXq9Lw8dOsS4\nuDhGRcXQbm9CRfEl0F8TpH7a+2ycMmUaTSYfyvEFKUcDdgKluGTJkgIcxhuhKD7XTUx9WKdOHTZr\n1oE63RMUYvjRtFjqcc6cOZnK1q9fX+s/l+B+RsB2S0cWly9f5qpVq/jPP/9kul6jRg1evHiR1atX\nT79WqVKlIj1RffPNN7TbGzHDanyeknSkPiURzT+U0wWT9sNXmEHYfpKqGnZH89H/F5CdnJD/vUXN\ng7yhOM4pxQHx8fG0Wv0o2UcHE3ieFov/DXO6O7788kuqagyFKaozxYJo1ZSjpyhWyCRKWuUYigtg\nIiUVsko5yZxMMUS4mDXO0mAw55iJNje4m+aUpKQklihRlhlZEWdTUV5hr179uH//fr7wwgSOGzee\ndes2JzCXEtQXqK1hUcxwi1hBRQnRWDDsFJea1wmUZcmSEfTzC2UGOcBRysanHcXK3IOSmMRJcbEx\nU9xqftdk6kVtHW2pPTtSKxNEnU697UyMuUVqair379/PU6dO3fTe0NByFPeg7OXk5hFcQCrES98d\nzrxI1u1AVVVMmTIRqloCwLMwmdYhIOAgOnfunGM5u92Ogwd34euvZ6NUKRVAU0hQQh04HM1Rp04T\nWCwW7N69CQsWDMGrr46GMMWMgHADKhg6tC+eeaY/gDRIgGwpAE0AOKEo59GiRYuCa3gWIB0Qh3oX\nAnDixCls3boJTud3AIIAJCMp6QB++WVjprKSiKUpxEkfEK7FhFwF8wHApk2bEBlZEV26PIMKFarg\n7bdnp39nNBrh65s5MDi3zy0sJCcnQ/aBLg5pLwB6yF4wBBLEdwkSwFIV4p7vImwPhV5/H/btyyqz\nqQfZoTjKiQeFA4+sFAyOHz+O1FQDJJFEaQBbkZpqwYEDB264d9euXfjiiy8QFRWFxx5rAbP5bQC/\nQdaZIZAkdj9C5suSkGD6sQDKQObL/pDkXSMga2tfCJnEW1DV5hg+fASMxlzz7WeJu0lOzGYzFi/+\nAFarCcBAKMoJ2GxvYvToIahQoQJeeulFvPzyZLz++kSo6jjI2tUcJpMRNttV6HTvAZgGk+kJ1K1b\nBjpdS0hCmTkAZgAoiytXyqJixXvg5dUXMnaVANQD8DOEwOAiZI0sA4NhAMzmagC+h6iGEyGBgTEQ\nDu4akKC/dgB+hE7niy1btmDx4sX48ccf4XA40tuWmpqK55+fhBo1YvHgg92zlMeb4dixYyhXrhpq\n1myNqKhoPP30ELz99juoXbsZYmPb4bfffku/NyUlBadOHQHQ5Zbfcz3mQUIn/wRQHsJOPSfPT705\nstT+v/76a/btO4gTJ77ICxcu3NIuIyiorLaz/VvbJTmpKI35ySefZLpv06ZNrFPnPsbE1OKMGTNI\nkr///jtttoraDu4Tin/Wx/TxCbmj/skkKT5eNSlpQRcSUFmvXj1arSHMSC9KAoPYunVm/y85Pgtl\nBp3e5wRsufZRDggId7OoHqGqBqfzV/bu3ZuffvopK1euzAMHDnDQoEHs169fkd7Rnzt3jv7+YVoK\n8Y0UKiM/ig/3IQo35EVtN61qu/LvtPb/67Eo3waykxPyv2f98SBvKI5zSnHAgQMHtHUkni4XCCCE\nK1asyHTfG2+8TVUNod3elaoazgkTXuaJEyfo7V1CWwdslFO3VAr1W13N0lmSQt31AzN8lQdqZQbT\naPRnnz7P8LPPPsuX9fNunFM2bdrEZ54ZzuHDR3L//v1Z3rN161YOGvQshw4dwb/++otHjx7lyJFj\nOWDAUK5fv55t2nTS1rvm2r8JmjwkUVUjWL++ixrunHb9Dc1C3J5AA3p7l+SaNWs0goD3KKcGBk0G\nGlFivny0tZQEfqfB4KW5/HSil1c1NmnSNv1Eu1evfrRamxP4iTrddPr6huTKKuyO++5rpQWpCqWh\n0RhBszlaW7fnUVUDuWPHjvT7g4PLULio8yYnNgCvQrLxbQHwCgqB9SI/0LVrL4p/1FU3ZbI7u3Xr\nxpUrV6Yr3mlpaVy7di1XrlzJ8+fPkyQPHjyouV7YNYFSCZi5dOnSfK/nzSBHV4MpRypNCXRn2bLl\nabUGM8M1hAReZfv2XTKV7d+/PzMc+CtRjmQsueJRzo71YuHChSTJq1evcuzYsaxVqxZr1arF559/\nvlhEqO/fv59NmrRjaGi0NvGXo/jWracwrZyl+HtVpAT32aiqVWix+POll6bm+f13G7KTE/K/u6h5\ncHsornNKUcdff/2lzf3ugcrRXLlyZfo9Z8+epdnswww/1DO0WAK5bt06TpkyhVarr7aenqb4IQdo\n66I3xW2xAcWFrZ+2bvpTjuAjqKrB+doez5xyewgLq0gx/O0iUCGTPHh7V2XZsjUpsTkSvC+5JepT\nCBFsHDtW8ihUr15fG/eamhy8SkUJotUaToPBR5OLcgRUTYdwGdtSabM14ieffEKHw0GDwcwMbm9S\nVR/iBx98cEtt8vFxJV5xyXU4xR/b9fcEPvfc6PT7N27cSB+f4IKWkwLDDR1w7tw5tmrVhb6+oYyO\nrsONGzfeUgf++++/miLUnUJj8ok2cGXp7d2MAQHh3LVrFxs0aEEvryr09m5Of/8w7tmzh+fOnaPB\nYKdYcUlgDY1G73yhtrlVyKSz2W3gB9Db24c2W7AmxNsIfEsggJ06dc9UVjLz9aJQy22i+CrrcqUo\nOxwO2u2BlGQt4l+mqhHcvHnzTepbfCaqt956i0KV46v9sJsxg3bQl4CdihLKadOm8fjx4/n6bg+K\nl6x4UHjwyEnesGfPHi0eZwQlUcRrVBQfzp07lzVqNKavbyhr1GhEm61cJsOI1VqeZrM/vb3b0GoN\nZ3BwOU3B6kUJ8LpKoQp7mBnsCQ0pWeNaaPPoTAYG5m9mvpzwX5eVbdu2sXLl+vT1DWWzZh14+vTp\nXJf18wvXdIFkiuFsLIHd1OsnMyysAi0W1zo4Vhu7hpQN1usExnDcuBdIksuXL6fVGkrgHQKrqKrV\n+Nxzo7h7926eOnWKdnsJig/zJ5qO8li6TBkMIzhlyhQ6nU4ajSpl4+ViyOh40zi061GjRiMqymy6\n2DkUpQTdMxTqdKM4evTzmcpcuXIlRznJKeHIcq2gksV3hLBh3FE88EBX7NxZDampv+PSpQ1o3rwd\n9u3bjtDQ0FyVDwkJQVRUFI4ePQPxzU0FUBkpKb8hJUUPRXkLHTp0x+nTpZGYuB2AHoryLp54YjDe\nfPMlqGpFXLnSQHtaLHS6YBw4cAC1a9cumAZnizQISXgDACmQpEF6lCoVib17S0GSpKjQ62uiatVK\nmUq2adMGU6Y8iORkJyS5SgpiYmrnyp9Lp9Phq68WolOnLjAYKiIl5QCGDXsGL774IhRFAXmjnClK\nVuJTNHHx4kXcc889EIKX6hCSdQOA+RC/KxXir7UZAwcOhM1mK6yqFku0a9cuRzn53//uBJmOB8UB\nN5MVD/KGwMBAGI0OpKQchPhnloHJZMXo0S/i8uXxINtg1675cDrfALAMEpexComJxwD8juTkqgBe\nQFLSe9DpAKezPyS2wwbxSR4O4GEAxyHJLv6GHESvANALHTs+mi/tuNvnlLNnzyI29gFcuTINQCzW\nrZuFFi06YufO37L9nfz666944425cDicCA0tgStXhsPhmA5gEBRlHHx9P0PNmjUwadKHaNXqYQBP\nQ9a/XQDKQRLIVANQDaraFADQtm1bzJ07DePHT8HZs+fg5+cLb287YmJisHjxYly7Vhninw4AHSAx\nUu8DOAOT6UvUrTsfiqJg8OAhmDOnLRIShsFg2AEvr53o0GHeLfXJZ5/NRaNGrZCaugAOx2mUK1ca\nBw8+joSECVCUU1DV+ejTZ0OmMna7/Zbe4Y5zEKaLUQDu1z6x2uf+235q7pFJ4798+TINBlXbobqO\n/Tty8eLFOe4u0tLSOGbMBIaFVWS5cjU5c+ZMRkbG0GSya9Geb7jtmP+kzVaCmVNT72PJkmV5/Phx\nWiz+zDiGOkRAZdOmD9zSbic/IJHGNgoFTzCBQJYrV4GrV6+mqgbSYBhGi+UhhoaWuyEJSlxcHO32\nkpol4XMqSiX26TPwlt5/5swZrlmzhgcOHCBJBgYGsnr16pw6dSrXrl3LtWvXcs2aNVyzZg3Xrl1b\nLHb027Zto69vCH18GlL8kMcRGKVZl6trfe5HoBMNBr98eefdhpvJCfnft/54kDv8F+aUoo7hw8fQ\nZoumTjeKNlttxsa2ore3e1pqJy2WCPr4lKTJ5E0vrwCaTP7adx0I1KLEuMRQ/FidBGZQqEt9KH7K\ngZR01q5nXiOgvyEu6HZxt88py5Yto7f3A5nGzGz249mzZ7O8/5dffqGquqhl6xAIpqJUobhG3EOD\nIZCDBo0gKe4sFosPgQEUt5nKBN7UxrkaAWsmlpTTp0/TZvOnuGA0IODN2NjWXLRoEYHGbnW8QkBP\nRbHQaLTy9dffTH+G0+nkrFnvsk2bHnz66cH8999/b6tfLl++zHXr1nH79u10Op388sslbN/+Efbs\n2Zd//fVXlmVuV04MkHyTCyAK88uQ0Mc7hUyNSE5OpsFgoXAgC4+fl1ctfvfddzl22NixE6mq91E4\nHFfRag3mmjVreOHCBX744YdU1doUnxgHjcZnWKtWA9ps1SnBbg4ajUP44IPivtC6tcsxvZU2GbxF\nQJ9tlpmCgqL4aYJOypHJvWzZsiVJ8tdff2W3bt3Yt2/fLDMFSpvdSeNP0WTKXTAfKe4Xs2bNZufO\nvThy5PO8ePEiU1NT+e233/Kxxx5j9erVOW7cOO7evTu9DIrBRFW+fA3KsZBTU5QDtYVgLcVP2YeS\nxSiCAQER+fLOuw03kxOyeMiKBwWP/8KcUtThdDo5e/ZsdujQgVOmTOHWrVtps7k4dkngIs1mX546\ndYrnz59nSkoKAwMjtHlST+Cydt8/2vwYSUlmsZHALxS+3kmaMee4du90ApVZrVrjfGnD3T6nrF69\nml5eVSjBlOJHbjBYs+Wlbt/+EUpytQ8pwXausZ6tKbMXabOVS99kDBw4mOI286s2rjEExlCMhG1I\nSuzSL7/8wr59+1Lod//QnnmSgDc3bdpEk8mPwEiKb3IsDYZgfvTRR3mmBcwtkpOTuWXLFu7cuZPn\nz5/nyJHPs3PnXnznnXfT3U7zQ07MkGTtcQAG5fVhucQNjX3xxVepquUJTKKqtmbduk1vyv9bqlRl\nZs5iM5UDBw4jKRNFv35DaTTaaLEEsGbNRjx9+jSjoippypKdBoM/t2zZQpIcOHAgxb9mOSVL4DUC\nBp47dy6fhzVnSDDfUbc2TWbdug3422+/Ua/31iarQAYFlUkPaHDh/fffp8Hgriifp8FgybWi3Lfv\nYKpqPQIf0mR6kuXLV2NCQkL690lJSZw/fz4DAgI4c+ZMrb5Ff6KyWLzTN0fCCalqirI/hUtyHCWA\n4SEOGjQoX955NyMrOSGLh6x4cGdRXOeUoo5Ro8ZQTsqiCah85JFebNu2O1W1IYEXabNVZ//+wzKV\n2bJli8ava9DWP9c60khTiFe6XfuIwEME+mjraZCmaH3EihXr5nt77sY5JS0tjY0ataaqNicwiTZb\nNMeOnZjt/W3a9NCU5OcpXMeusfqHwoZF2myPpwfQtWzZle5Z9WQt9GX37o8wOTmZP/30E1U1gHp9\nEOUEIUJbN115Gupw5cqV3LFjB729w6go/lQUK/v2HXTH2MLOnj3L8uWr08srmqoaSYslkCZTbwIf\nUlXr8emnh5DMm5xYIA5MXwLYDGA8gLD8ka+bIstGL1++nGPHjuPs2bOZnJx8006qWLEOxSIoA63X\nD+OYMeMy3XPx4kWeOnWKTqeTc+fOpcVSh5L+sTcVpTcbN5ad05o1ayguDy8SWEWgJc3mwDtOD2ez\nhWpKmytFZAUOHDiQPj6l3IQ/iUB9PvTQQ5nKyjGIjeJe8gOBRlTVkrl6b0JCgmbVv0TXMY/d3pD/\n+9//mJiYyCVLlrBr166sXbs2J0+enJ7iGcVgoqpevSGBaRSSfLv2Lym7aD9KsMoYAgH86quv8uWd\ndyNykhOyeMiKB3cGxX1OKco4efIkxfq3RZvnDhKwcf369Zw3bx7HjHmeixYtynJtS0lJ0VwwWmlr\nyCRtTWlM4AM3pWoKgd4UNw1VU6ZHEQhljRr1860td/uckpKSwrlz53Ls2HFctmxZjvd+//33VNUQ\nAkMowXsXNT1iAiXz3jGqahg3bdpEkuza9XEqynS3MZ3DZs06khRrvgTp9SLQjRLM6aS4avQlsIeA\njY880pOHDh2izeZPvb4sjcaq9PUN4aFDhwq8b0iye/cnaDQO0+q2lMIW5mL3uEi93pwnJp1PIFn5\nXgZQJT8lK5fIl05avnw5VbUkgZep1w+hn19ojmwFzZq1pFgRy1MyyYTQbg9K/378+Bep0/lQpytB\niyXglpk38gPe3oEUq3IgxSLgpbFZ+BD4y02oZzA8vHymsjNnzqTJ9JDWtqbaJKfkivUiPj6eBoOV\n4u7h8hN/gI0bN2aNGjU4bty4dE5ld6AYTFRPPtlfG3cfCsUN3T6hWj+bCJShwRDI6Oj6HDZs9B13\nuynO6NmzZ45yQhYPWfGg4HEzWfHISd6wdOlSypG6+zxXnffcUy1XbD6///47LRZ/jVHArhloftPW\npBcoMTCuWBpvAo9SXBUjCBhYvnztfGmHZ07JHocOHWLr1l0ZE1OfzzzzXPrJ74oVK1ivXiuN9cTG\njHTTgTSZvNmt2yOsUSOW997bnDNnztTSYo8hMJ6qGsg//viDpPgkWywBFBaxz9zk6GcKQ4aFwGM0\nGOrRag3SNlIzKZzN1dm+/cMkyUuXLvHxx/szJqY+O3Xqedt+ydkhOroexRWIBJZQ2FdcdU2mwWBl\nfHz8bcuJE5JeJavPlfySshyQbx21YcMGDh06gi+8MCHHSWDVqlXU6/21zlxJcWEYTYslM+fjhQsX\nuH///kJTkkRJrk6xcI4mYGdYWLjmu/y8tltKIHAvy5e/J1PZ1atX02YrywwC8U8YFVUp1+9u2bIj\nLZbuBH6lTjeVgYERVBSFXl5eWX7sdnuxmKgqVqxLSdZSTps8XGk+f6VYQ97Uds7eFEvJOlqtbdmt\nW698ef/dgJvJCXn3LmoeZMZ/YU4pyli/fr2myPyuzXN7CfhRUQYzNLQc4+Pjb/qMpKQk7tmzhxUq\nVKfROITABhoMj2nP9aLwJ9fWFCOXYnKAgJXduj2eL+3wzClZIy4ujgEB4dTpXiOwjhZLJz74YLf0\n7/fv308vrzIEzlCICa7Sy6sWhwwZSlUtq+k/S2i1BnPevHkcOXIsn3tudKbNSFpaGn18SmrrYidK\n0hoHxdUmkOLKmEJx1/Bnhj90EoFgVq5cj06nk/feG0uz+UkC62kwjGGpUhXzNcX1ww/3odH4jKYX\nnSbgQ0WZQuBXWizd2bKlWMgLWE4KDPnWUTkhJSWFq1at4rJly9i27UME5lB8q8pSssyUZIkSZe5I\nXXILUdxCKQGFvgSq0GKxsXTpGG2CCqNYRn05YcLEG8qPHPkCjUY7rdYI+voGZ8pSczNcu3aN/foN\nZcWKddmqVZdcHZ+gGExU1as30PrSn+J/ZyEQRbGKlKa4ZaRRdt+7tB/8Ver1pjsWkHA3oDjIigeF\nD4+c5A3nzp3T5jkrJQjPj8DHBEhv70b84Ycfsi178OBBLlmyJP14Pi4ujj169OY999TRknr5Uk5k\nXdzzVbS/Qyn+q7o7GtdzN8rKokWLaLe3ddugJFKvN6croHFxcTSbvQn8SxcThcVSknZ7OIUH+yVK\ngOC77Nr18Wzfs27dOnp5laBeX0JbO8O08S6nbZBcOko4M9wdnAQi2bXrQ2zatD0VxcXL7NDk716u\nW7cu3/oiLi6OMTH30surLFU1jPXqNWGLFp1YsWJd9us3NL1PcpKTnHiU//NITExEw4atcODANeh0\nJZCYuBmAFcAiSM7zUADD4O19JxIR3goUAF4AXgfwD4DRUFULQkLCceRIIwAtAJgAjEVwcMlMJUni\n9983Ii3NB6mpZZGSshVHjx5FtWrVcvVmVVUxZ86b+dqawkZycjIOHNgH4QC1Q/i10wCcBzAWQC0I\nj+QZAInI+NkkQFF0ueKg9sADDzwoKlBVFSaTDSkprwF4DsBeAKUAEOQ1GI3GLMt9/vkiPPXUYBgM\nDZCWth19+vTA229PwxdfZHDdKooKift/B8ARiPfmRIgH53gAKi5duoTAwMCCbOJdDaPRCPKa25Uk\nAIRerwcABAQEYNy45/Haa/VBtoairIbDkYL4+AEA6kB0ixMAysNkulFNvHDhAj766CPs3LkLXl4+\ncDguwmr1wuXL98HheBJAJwDPA+gHYACASwBKAugNgFDVBPzww2pcvToa5NMAJgO4CGAynM4EGAz5\np5oGBARg587fsH//fhgMBlSoUOE/xcWebzuK7DB16jRaLB2Zwc08iuKHOsZtJ7aJer1/gdflViA+\nYTvd6jiMUVGlGRxcntf7KPfrNyRT2dmzZ2u7vUTtnh9pNPoWcH2L9o7+0KFD1OsDCQzTLCsjCMQS\neNKtL49rlnwbgXYE5lNVa3Po0FH50EMeuFDUZcWDogGPnOQNf/31F+328trc1otAEwIf0Wh8lJUq\n1ckyUD45OVljB3KtPRepqqVuyMwqc+kC7R5XPEwGHSlgYpky1e5UU+9KWbl69SpLl65Ek6kfgY+o\nqvXT2R3csX79es6aNYsjRoygzdbFbZwuETDSag3g1q1bM5WJi4tjSEhZGo3tKSfXKwmcpF7/JM3m\nEjQYQig0gC46wNoUV8adBCJYpUodTpw4kSbT027v+5uAHy2Wh1mjRsMcT2mvXbvG1atXc926dUxM\nTOTrr7/JJk06sGfPvjx27Nht91lOcnJXW5QPHTqGpKT7Abgsgo9Br58Fh8MBsSqmAEgDC/Jndttw\nZPp/XNx5VK1aF6dPD4Hs4lUYjXpER/fOVOqHH34A0BhCaAIA9yM19TLS0tLydRdXnBAcHAyn0wmg\nAoBIAH4Qi3KS212pkMxTybBaf0HLlt5o3bov+vXre+cr7IEHHniQB4SGhiIl5TSAZgAOATBCpxuO\nJ554GDNm/ASTyXRDmQsXLkBOKqtqV3xhMFTDsWPHMmWndTickPkTkLU1Wfs7Mf36P/+cLJB2eSCw\n2WzYsmU9XnppKo4c+RHNmvXEM8/0v+G+Ro0aoVGjRvj0008BHHT7xgFFAdat+w41a9bMVGbu3Pdw\n/nwjpKbWAxAIoI2UcLwDh+NjVK1aG7t2ucZ/BYApkLUVACajdOkf4evrC0U54fbUNBgMDgwbVg7D\nh78JZqN0nT59GnXrNsHFiz4AkmE0XkBiYkkkJo6CXr8D339/H/bu3Zar04pTp06hV6+B2LFjO0qX\nLnvT+4sqbntnkFssWLCAqlqD4nSeRpOpH0uXrsgMhgMzAT/6+gbf/GF3EGLZLKn5e9UkoNLHx5dP\nPjmAkm1nD4GfqdP588cff8xUtkOHDhTfsb81X6HpBLwKlOIOxWBH36pVB2YkknmE4r9n1Kzv71G4\nRktRIsXVW/Lr9iD3KA6y4kHhwyMneUNqaiqDgkpTgsEPEZhBmy2IV65cybaMw+FgUFAkM3h1d9Bq\nDbwhTkVOPO2aJTFGW0ct2sefQMk7ekrrkZWb48KFCwwKiqLBMJrAYqpqAw4YMCzLe0eMGENgMiUr\nYxNm+B7/RMBORfGhXq8SGKzpJ++6WY6fZ+/eA3jy5En6+oZQp5tM4AuqalWOHTuezZq1p8FgocFg\n5rPPjrlBL+nW7XGtji5f5+4Uqjt5vs3WhfPmzcu2nadOneK6det4+PBhVqhQgwbDGAKHqCgzc5ST\nu9q5smfPnnjyyaYwGMJhMvmjevUDKFMmEkAIxPf3KoD2SEhw5PygO45UyNA9BaA1ACI8PBzffvsz\ngPcAxABoCqdzJJYtW5mpZKtWrbTy1SGW08nIbDm9O7F06UI0agQAFwCsBvA3pF9aQvyUDwO4DLGO\n6LFt27ZCqqkHHnjgQd5w5MgRXL6cArH2lQXwLByOcGzfvj3bMjqdDj/8sAxBQeNgsQTCYmmMefPe\nQdmyma1xYox2AngYQHNI3M+fABIg/qoJiIgILohmeXCb8PPzw/btG/DYY1fQrNnnmDy5O2bNmpHl\nve3bt4GqvgugBGRNbAFgCIAOAAaCfAMORyCArwDsBzBK+/5JGAyzMH78SISGhmLLll/w0EPH0Lz5\nYsyYMQgnTpzBr7/akZZ2BWlpJ/DGG0vRuHFT7bRXsH//30hLa6X9pQB4EKKrCUhDttbopUu/Rtmy\nldG+/VhER9fC33+fRFpaDIAfQLbIQ+8VLvJzw5Qjrl69yri4ODqdTjZt2orAa247oH00GPzuWF1y\ngwxmC3261TMmJoZWawgzZ0bqz1atHshUdtOmTZpF+mNKWu8HGRVVuYDrWzx29HFxcVQUI4Viz9WH\nJ7T+Uij+yV0I3MeyZavm23s9yEBxkRUPChceOckbDhw4oM1rrjTUyQSCuXz58puWdTgcPHXqVLYJ\nv4zGQAJltPVJoSSfcM2n1wjo2aNHz/xuUrbwyEr+4/PPv2BoaAV6ewdTUSwE6hLoSeBptxMEP0q6\nc7t2Qqtnhw7duXr1avbuPYADBgzlvn370p8ZFVWVmbMozyRQkoMGDU+/5+mnh9Bs7kmhtWupyZiZ\nwCDqdK/Q3z+MZ86cuaG+V69epar6MSPBzq+aHtWMwFMUOjuPRTlH2Gw2BAQEQFEUXLoUB2AdMvps\nA5zOohYhqYP4ll2DRCvH49ix4zCbFUim8RchO/clKFEiNFPJbdu2wWLpCqAXgGoAPseJEwcz7dru\nVhw/fhxmcwkA6yEWEQD4HYBR+3cBgO8BtMfRo/9k/RAPPPDAgyIOiUnxne72+AAAIABJREFUgqwj\nrwFoDYNBD4vl5gxPOp0OwcHBWfoxA0BqaiKAAADnICwXv0NOMaH93wt//nk4743woNDw0EM9cPLk\nfly+fArLli2G0bgLYj3+G8IOdRxABIRBrCJEFs7i+++PoWXLzpg/vzzmzPFB7dqNsG/fPgBAWFgo\ngN+0NxDAHwDaY+HCZSCJzz//HBs3/gHye4jPczjklGIXdLrFaNBgDTZvXo+goKAb6vvvv/9CUXwg\nDFYAsB3AfQB+BPA+gI9ybK9HUb4OyclOiFP7fZDs3WPgdKYUbqVugAOiDJshx2YDkJycijp16kGv\nbwNxGSgBq7U07r+/XqaSVqsVev15tytxMBjM/ym6lNtFVFQU5Ie3HUBdAN0ADIQcLf0GoDOAKADf\nw2wuapSBHnjggQe5Q6lSpWAypQFoBaHBbAyDIREVK1bMh6ebAYyDuPb5QJTk2gB6AHgIwFWoqjUf\n3uNBUUD79u2wY8cWKMrfAEZCxjwQ4rK4HcAEAL4A/JGcPBJpaeUBXAR5FVevPoi33poDAHjvvRkA\nRgN4AMD9ECNgU5jNJnTs+DAefXQ8du6MQkoKIDrQK5Dg0gogn0Lz5o1QpkyZLOsYHh4ORbkKMYIC\nwF8Qpdml91TNspwLHkX5OlitBsgOaCxEUZoEnS7rnXPhQQ/JLg7Izus3qKoFH300C1FR2+HltRgW\ny7to374SnnzyyUwlO3fujBIljsBkehLAm1DV1pgw4QWPogzA19cXL7wwEtKn3SCK8SYApwEEQxaU\nfwD8gRkzXiy0enrggQce5AU2mw1Lly6EzTYHdvtKWCxv4f33ZyI8PDwfnp4GYKP2/1AA3hCrdQcI\nt7IFr7wyOh/e40FRQUxMDBo3rgNZL13YCJ3uGnS6rW7XNkFYVi5BfJxXYN++venPGDt2OCRGyB/C\nxdwPjz/eCT/+uAvkHoiF+meI6up6LmE2b8XXX69EzZpNMHLkC0hOTs5UP6vViqVLF8LLqyvs9hgY\njQtgMs2HKMwJMJkm5Ni+oqwdkYXAyzZgwBDMmbMYQhMWBuAn+Pl54cKFf+94XbKDotgg9G4PAjgG\n4BAqVfLD7t1/IjU1FYcPH4aqqihVqlSW5S9evIi33pqJkyfPoXXrWHTp0qWA66sABSdreZKThIQE\nxMXFITQ0FAaDAePGjcOrr86AJB9pDmAPgH8hk/x6ANfw3HM98frrr+dH3T24DkVZVjwoOvDISf4g\nPj4eR48eRXh4OPz8/PLlmRZLCYie0gjiwrYWckweA+A7AAkg71yAvEdW7gwOHz6MOnXuR3JyfZDJ\nsFp34OuvP0OnTo8gKakOACdSUtYgNbUdgE+0Uqvg5fUk9HoHAAVDhw5EZGQo3nzzXRiNRkycOAaX\nLl3CgAHfIiHhC60MIacWRlitnaHTHUFS0h44ncNA3geDYTqioy/hl19WwcfHJ1MdXfIeFhaGxYu/\nwsiRY5GYGI9mzR7EqlVfA9nIiUdRvg7Lli3DI4+MRGLiswB0MJl+RMeOdixaNP+O1yU7KIodwDIA\nRyG79U2oUGEF9u/fW6j1yg5FdaL64IP5GDRoKPR6O2w2Iz799D20afMIHA49gCaQ48KJAJZDdsEV\nAYzBggUD8Nhjj+VX/T1wQ1GVFQ+KFjxyUnRhMpVEauoUSGwHIfkIZgJ4BjrdVtx772H88cdPd6w+\nHlnJPyQmJmLDhg0gifvuuw+qqmb6/ty5c/j222+h1+vx4IMPws/PD3Fxcfj222+hKAp++20j5swJ\nBvCCVmIXZK3dCMnY1w3Tpw/AwIH90p+5b98+1KrVGAkJ3wGoAckaOB2KkgKz2Qlf3xI4exZwOj8A\n0BTiPumDyMhy2LVrI7y9vXNsE0koilLQclJgyNcozVvB5MlTaDRaqdMZ2bRpO16+fLnQ6pIVJFq5\nnsZa8T8CXrTb7YVdrWyBIhh1vHv3bqpqSUrGIBL4hKoaQKASgWkEKhLQaZGxTxHYrUXhqpw6dWo+\n95AHLhRFWfGg6MEjJ0UXwkIQpjELbCAQTkBPnc7I6tUb8t9//73D9fHISn4gLi6OpUtXpt1ej3Z7\nfUZFVeLZs2dv6RkbNmzQ1t2VBLZTp6vNzJkbv2STJh1uKPfll0tos/lTUQwEvDUmlaWUfBKfEJhP\nIIjAagIXCZhosXTi7Nmzc123nOTE46OcBcaPH4OkpKtITLyKn3/+3013JHceBkhU8cMAXgYQiZSU\n1JyLeJAJO3fuhF5/PzIyBvVEYmIyJIjgVUj/BkICBhIhgZ1TAFTDn3/uKYQae+CBBx4UB5gBlAfw\nDIR9qRwAGxIS4rF9+y8ICQkp1Np5cHsYM2YSTpyIRXz8b4iP34CTJ5tj5MicfXuvR4MGDfDFF+/j\nnnsmITz8UZQpkwYgI/OfTncQJUrc6ALUtWsXxMfHYf78D2C3twIwHWKVfgtATwjb10uQE+C2AJ5G\nWloUrly5cnuNvQ53Z87iXECn02VLf1P4cAAYDvGf9QZwCD4+8wq3SsUMUVFRcDo3QwjTfQBs1Ijy\nv9L867pDmEUqQjaarQHsA/ATIiMjC6XOHnjggQdFHTqdCU7nExADAwFYodPthtlsLtyKeZAn7Nt3\nBKmpfeHyTkhNbYYDB2bf8nPatWuHdu3aAQD27t2LunVjkZx8BIADFsv/8OKLa3Hw4EGsWrUKRqMR\nXbp0SafvrVChAhyOSRAXi6yMgwcAPA2gHkymJ9C69Y+31dbr4VGUiy26QKh2jgH4DdWq1Snk+hQv\nNGjQAE880Rnz51eB0VgJqalbsGjR53jllf/DH388CMkk9DGkjx+GUNXsBHANHTrMLMSae+CBBx4U\nXfj6mnDhwmBIALQCYBm8vb0KuVYe5BWNGtXG1q0fIjGxNQDAav0QDRvWukmpnBEdHY0//9yEJUuW\ngCSOH++JypWrweHQQ6frArM5GePHT8GOHb8hJCQE9evXR6tWdfH119shvsoDID7wabBYXkDDhnWw\nbdtH8PH5BrNnL0C1atXy2uwij1vyfbmbAPgQWE7Jde4k0I41atQs7GplCxRhH7Ht27dz+fLlPHbs\nGEmyR4/eBN7W+jVc87Fz5ZVvQEWpyXbtHsqPbvEgCxRlWfGg6MAjJ0UXqhpCYJLb+jSZqhpcaPXx\nyEr+ICkpiW3adKXJ5EOTyYetWnViYmJilvc6nc5bfv6iRYuoqjEEWhF4J91vWa8fxtjYltyxYwdJ\n8ueff6bNdq/2/QoCHako/ly0aNFN3+FwOLKtW05yUtA+yvMgaVr+dLvmD0mHcgDAKggTtQe3hDQA\nsyC+YP4AiLS0Yp1Zr9DkpHr16mjbti0iIiIQFxeH1NRrAEZA6PcuQHariRDLSC2QD2D58uWYNOmV\nbHPKe1Cg8MwpHuQWHlkpBDgcgCSaULXPFjgcRZpMwCMnuYDZbMbKlV/ixImDOH78AObO/T8cOXIE\naWlp6fesWLECAQHhMBiMqFmzMU6cOJHr5//8869ISHgKwBUAldOvOxxVsH79CTRo0AqffPIZ6tWr\nh4CAeBgMowGYYbFY0bBhHXTr1i3bZ8fHx6NNm24wmazw8grA22+/c0ttL2hFeT7EudMdYyACWAHC\nHD2mgOvwH4QCIduOhAzhr7h8+ULhVilvKHQ5mTVrDkJCIrF06UZIKs6RkH7upVXhMwCfQviVy2P6\n9M/x2WcLC7JKHmSNQpcVD4oNPLJSCEhNvQLJZOoPCYr+XbtWZOGRk1uAv78/Bg4cgYoV78W997ZF\ndHRt/H97dx4eVX3vcfw9M9lmMgmRJARCkURRQEAWIy7UgsUWklvxGhDE2lKpyyNV0FbE9V6qdHEp\nVx5wKa0I16oogpRSDCoQuTQalgKCKLiw70tYQkKGyfzuH78TGdCEk8kZzpnM9/U8PJmZzHzPd2Y+\nnDmTc87vt2fPHr744guGDbuNQ4dmEgpV8ckn/SksHGK6brt2rUlOXoGeVv0J9LTXm4EJhEJuqqrm\nc+edd5OSkkJ5+RIGD95Pjx6/Y+TINrz77tsNTpp2xx1jWLw4mdraQ1RVfczDDz9LSUlJE18Ja+Vx\n+je1z4Ec43Jr4/p3afSf7uOFHh5lgrHr4aiCbiopKdnutuqFuV1ftuVkw4YNyuvNUXCrgkkKPlSQ\nr2Cv8RpPM17zDsZwcWUKXlLDh//SgldHhHN6VoQzmMxJpFmx++nFNL2uvEnBSQVBBUMVpNvYj6xT\nrPTCCy8qn+8HCqoUhFRCwjhVVHSTmj59uvL7bwkb6i2kPJ5kdfz4cVN1jx49qjp27KVSU3+gXK7W\nxjCDLRU8aWRonPJ4klVlZaWpeqFQSG3dulXt2bNHZWW1V/BFWG8T1AMPjDvt/g3lxI7h4XLQuzkw\nfuY0cF/xnWrRQ6IApAE30QyPAjhnOdmwYQMJCZejpwVfCawHfgS0Mu7xM+AYenSMfsBVJCZ+Qm5u\ndrRaEo0j6xRhlmQl6hLQn08JgCfsckyRnNRj1ar1VFUVA17ARTD4U9auXU+rVq3QU0LXjUaxAfDw\nwx/eyKBBw9mwYUODddPS0li9ehkzZozmggtygdnAQfQwcMOBd7nooq6kpqaetcfDhw9z+eX96NSp\nN+3bd6KmJog+GR9AkZLyCa1bm//8tju9DW7Fjx8//pvL/fr1o1+/ftHvKCZ4gDnoIeJOAHPIzW1t\nb0thSktLKS0ttbJkVHNy0UUXUV39L+By9Fz0m4A96PnoM4B56ClYbwPG4fX+kJYtd/DQQ2WNWo74\ntljLirBHFHICDWRFchI5j+cktbVvAdcbt8zC4zl34/zLOiW6unW7GK/3n1RX3w0k4fH8nU6dLmbA\ngAFcffVfKCvrQzDYi2DwTdzuTpSX34vLtYnS0mtZv34F559/fr21vV4vgwcPZuHCD9m+/e8EAkXo\nl/4tWrY8TEnJUlM9/upXY1m3riOBwBL0NlIfkpJux+NZgNu9g7Zt93PJJb847b2zWx7f3qVRt1XX\nBtml0WiQaozIUKCgnYJu6qqr+tjdVr2IfNfXOctJfn5PBW8bh7JMV3C5crkyFHRWkK2g3Nhl85Aq\nKrpeHT582JLlitPFQlaE/UzmJNKs2P30Yprf305BR2Pd2VlBR+X3f8+2fkxmJZKcxGVWAoGAuu66\nG1Rqar5KT++p2ra9WG3dulUppVQwGFSzZ89WU6ZMMWa6/fqbwx2Sku5QEydONLWMiooK1aVLb+X3\nX6L8/k6qW7crGzVDcocOlyn4OOxQi6mqsLBYvfjii+rVV1/9zsNBGsqJHX9RngeMAJ4yfs61oYeY\nlpCQTDA4E32yWTrwEtnZ5s8ujRHnNCdFRdcydWoJJ08WAz8nMXE1RUX5/Otf5Rw48Dqgx6l2u2vp\n0aM7LVq0iGY7onFknSLMkqxEWUZGOpWVY4Ce6G2PtaSn/4/NXTWa5KQeiYmJLFw4h3Xr1lFdXU33\n7t3xer0AeDweiouLAXj44SfQh4lqLldtgyfchcvIyGD16mWsWbMGl8tFjx49SEgwv7l68cUXsHnz\nQmprrwBCpKS8z2WXdWPEiBHf9OokbwC70GNsbUfvu24JfMDZh10x/e0h3owaNVrB+QreUPB75XKl\nqo0bN9rdVr04+zd623NSUVGhOnW6TKWl9VR+f4HKzb1QjR07Vvl85yloq+B15XI9rfz+bLVp0yZL\nlim+LRayIuxnIidNyYrdTy+mzZo1S0Gagr8Y/9JNjXEbLbJOscfjjz+hfL7uCt5UbvcTKiOjjdq5\nc+dp99m8ebNaunSp2rdvn6XL3rZtm8rN7aDS069Sfn9XlZWVp9zuJOXxJKtBg25WBw8eVGVlZWrN\nmjXfjKtscp3iOJa+cM1JQcE1CpKMlZFPQaqaM2eO3W3VixgZ8P3EiRNq/vz5KifnfAUZCtooPblL\nJwXXKo8nQy1atMiy5Ylvi5WsCHtJTpzrvvvGKvAaIwT5FXjVffc9YFs/khV7hEIh9cILf1b9+9+o\nbrnll+rLL7887ff33/+gSkhIVV5vB+XzZaoFCxZYuvxjx46pRYsWqdGj71Ne71UKKhRUqeTk61Rq\naiuVnt5L+Xzt1cCBxerkyZMN5sTJo4Ar1QyHcrCCy+UHEtHTWG8BVtO1ay7r1q1r8HHnyrZt25g4\ncTKHD1dy8803UFhYCNHLmmU52bVrFx07FlBZmY0+se8f6N2Hm4Gf0aLFh8yc+RsGDjxzyE1hFWPX\nnOOzIuwlOXGu5OQsAoEAejxcF7CY5OQETpw4YEs/khXnmTp1Knfd9WvgZuBr4CA+33YqKvaQlJRk\n6bKuv/4W5s8vRI9eBdAH+A/gESCAz1fIs88OYdSoUVBPTuwe9UJExANMB25Afwm6gZ07nTECw44d\nO+je/UqOHbuV2trOzJp1h90tmXLo0CG6dOlFZWVr9GQuHvRwcdehR7xYz8mTn5Gbm2tnm0II4WiB\nQBC4Dz1pBMB/U1Mz2caOhNM8+ugf0cO/DUBvw/yEkyc3s3//ftq2bWvpsvLycklM/JiTJ+s2lDcD\nxcblJKqqCvn0000N1rBjHGXRZCGgu3HZBfSkffv2NvZzyrRpr1BZOZja2qeB0VRVxcbsdbNmzeL4\n8e8BBeiNZIBu6Ok0j5GQsIiRI4dz6aWX2tajEEI4WUVFBfozqVfYrb1ITrb2r4QitlVVHeH0bZhu\nuN0hYyxma40f/zC5uUvw+6/D77+exMQAHs/f0Bvox/H53qFnz64N1pAN5RjUt+81wDigCvgcl+tF\nnnrqDzZ3pVVVnSAYPC/slvPqva+TVFdXEwq1R5/sXI4+r+MhIJVLLsnkgw/eZvLkZ2ztUQghnKym\npgaPpxZ4BtiHnob4CYqLf2xvY8JR+vXrj8fzCFCNnuDrLzzzzB9ITEy0fFmZmZmsX7+cGTNG8fLL\nP+Pf/y6lfft5+P0XkZKSx6BBHbntttsarCEbyjGoR48uwEL00HA9gCMNDuJ9Lg0dOhif7yVgFvAR\nPt+ddrd0Vkoptm3bQW1tCTAUPVC+l4SEGZSUvEV+/oUUFRXTtm1HFixYYHO3QgjhTDk5OeTn5wNb\n0YestQU2UlQkG8rilNdem0rfvgdwu1vg9f6AyZOf5N577210HaUUe/fu5dChQw3ez+/3U1xczNCh\nQ+natSuff76K8vJ5fP75St54Yxpud8ObwnIyXwxyuzNQKgUIor+ReRg4sA/vvvuuzZ3p4A4ePJx3\n3lmIy+UiKyud/fu3gkNPpjh69CgDBhRTXl6OUrVAMnrlfjXwIbm5cODANQQCvwPW4/MNZ/nyJXTp\n0sWa7sVp5MQbYYbkxLkyMnI4ciQE1Bi3pJCVlcL+/dts6Uey4lyhUOisG6n1qayspLBwCCtWLEep\nIDfeOJjXXvsrHo/n7A/+Dg3lRP6iHIP0f8wR6N1aW4HWrFixyt6mDG+99RbvvfcZ8BVKHaSiYojd\nLTXonnvGsnr1+Sh1GL2rsDNwNzAVeIpdu74iEHgOyAH6EwoNYfHixTZ2LIQQzhQMBjly5AR6gqb9\nxr/eHDhw1N7GhCNFupEMMGbMQ6xY0Yqamn0EArv5xz+28Nxz0TlpVDaUY1ItMAr95ScLuJXU1FR7\nWzKUl6/i+PGb0eO1uwgG77a7pQZ99NEqamruQp/A50ePNb+y7rfogWGWG9cVHs8XZGTUN/a8EELE\nr927d6PXmXeh984lA3fidlt/7KmIb/qz+0503lKpqhrBsmXR+YOhbCjHoMzMbOB941oQKOGuu5wx\nDFuHDnl4vaXovsDlWmRrP2dz4YV5eDx1PYaABcCnwJXAn4GOwPW43cPw+X5Cfv4xbrrpJpu6FUII\n58rOzsblqkGvR+vmcCihV6/O9jYmmp0OHcI/uxXJyYvp2DEvKsuSDeUYlJ2diR714kfApbjdW8jK\nyrK5K+3222+noMCN39+D9PTryMj4rd0tNeill54lK+uvpKf3Iy2tgDZt1uN2r0UPgv4VsBb4J4mJ\nC5k4cRDLly8hJSXF3qaFEMKBUlJS+P73+wDvoM/z6APMoXPnjvY2Jpqd559/mlatXiU9vS9paQV0\n6LCRRx99MCrLkglHYtDevXuARejp6dMJhRbx9ddb7G3KkJSUxJIl8ykrK6OyspIrrriCzMxMu9uq\nV15eHps2reGjjz4iKSmJPn36MGXKFB57rIzq6rq++xEKBRk2bBher9fWfoUQwslCoQRgCnpoUAUc\n4auvptnblGh22rVrx8aNqykrK/vms9vqWf3qyIZyDOra9VLKyuZRW/s4cJTU1NH06BGdb1KR8Hg8\nXHPNNXa3YVp6ejoDBgz45nr//v157LFn0NOD5wGzycjIpEWLFvY0KIQQMeKKK7qzatUsTpzQk02l\npPyU3r1loiZhvbS0tNM+u6NFhoeLQdu3b6dv3yL27askGDzMiBE/56WXnqsb3sRxYnF4nkmTnmfc\nuEdITGxFYmIV7733dwoKCixfjjhdLGZFnHuSE+eqqqpi4MDBrFy5BoCCgu6UlMzB5/PZ0o9kRZgh\nw8M1M8nJyQQCQY4f300gEKS2ttqxG8mx5tChQ/TtW8T99z9ATU0Ij+cor7/+smwkCyGECSkpKQQC\n1VRXH6G6+gg1NdVyXodoskAgwMiRo0hNbcl55+UyadKUc7bsuNlQLi0tdXxNs/V69+7Pzp0dge0o\nVcrLL89m0qRJTappVjReRycZMmQES5e6UGonMJ8jR4LceONwNm/eHHHN5pQ9u2s6SSy8ZvHaY1M5\n/Tk6+X0dMuQWysv3Ay8DG1i+/CDFxcMsqe3ErFjJye9rtOqZrfngg//FzJlfU1X1KYcPv88jjzzH\n3LlzI67XGLKh7KCaZutt27YVmAhkA5cB9/Dmm283qaZZzXlFpZRi6dKF6GnBWwJ9gaGEQhewbNmy\niOs2p+zZXdNJYuE1i9cem8rpz9HJ7+uiReXABGAj+hyPCSxZsrLBx5jlxKxYycnva7Tqma05b14J\n1dUTgDZAF6qqxjB37sKI6zVG3GwoNyceTxJ6JVRnPdnZLe1qp9lwuVykprYEDhi3KGAjLtchWraU\n11cIIc4mLc0LbAi75TP8fjn0QjSN/gz+/JvriYkbyck5N5/LsqEcg8aPvx8YCtwDFJGQsJjnn3/e\n5q6ahylT/oTb/SowGrgWt/tTCgouZuDAgXa3JoQQjvfKK5OA36PHUh4BTGDatOfsbUrEvMmTf0dq\n6v0kJt6D1zuczMwF/OY3Y87Jsp18BtgaoLvdTQhLfAj0i1JtyUnzIlkRZkQzJ6Xo465E8yDrFGFG\nNHMihBBCCCGEEEIIIYQQQgghhBBCCCGcYyD61MYvgHEW1dwCfAKsBpZH8PhpwF5gXdhtLYH3gU3A\ne0CGBTXHAzuMPlejXwuz2gFLgE+B9eiz0ZrSZ331mtKj1eIhK07PSUM1m9Kn1azOyhaclZP6ao7H\nOVmJx5yA87JidU5AsiLrFHPi9fPHUh7gS/Tgi4noA+U7W1B3M/qNiNQ1QE9OD8vTwIPG5XHAHy2o\n+d/AryPssTV6AGAAP3r8uM5N6LO+ek3p0UrxkhWn56Shms05K07LSX01nZSVeMwJOC8rVucEJCuy\nTjEnJj9/nDY8XG90+LYAJ4GZwA0W1W7KCB//B1SccdsgYIZxeQbwnxbUhMj73IP+zwpQCXwGtCXy\nPuur15QerRQvWXF6Thqq2ZQ+rRStrDgpJ/XVBOdkJV5zAs7KitU5AcmKrFPMicnPH6dtKLcFtodd\n38GpJ9wUCvgAWAncYUE9gBz0LgmMnzkW1b0XWIue/7Oxu0nq5KG/BZZjTZ919T62sMemivesODEn\n4TWbe1ZiJSfgzKzU1WvuOYHYyYpV70EekpVIxEpOwJnrlPCalmbFaRvKKkp1+6BfvELgV+jdCVZS\nWNP7i0A+ejfCbuBPEdTwA7OBMcCxM34XSZ9+4G2jXqVFPVohnrPixJzU1YyXrMRCTsCZWYmnnEBs\nZMWq90CyErlYyAk4c51SVzMqWXHahvJO9IHZddqhv6k11W7j5370dEG9Lai5F31sDOjJx/dZUHMf\np0LyVxrfZyI6fK8Cc43bmtJnXb2/hdVrao9WieesOC0n4TXjJSuxkBNwXlbiLScQG1mx4j2QrDRN\nLOQEnLdOCa8Zlaw4bUN5JXAR+s/nScAwYF4Ta/qANONyKvBjTj8wPVLz0PNzYvyc28B9zWoTdvlG\nGtenC717YQMQPl9opH3WV68pPVopnrPipJw0VLO5ZiVWcgLOykq85QRiJytNfQ8kK7JOMSMeP3+i\nohB91uKXwMMW1MtHH+i9Bj10SCQ13wB2AQH0cUm3oc9O/YDIh105s+ZI4H/Rw8OsRQelMcfpfB8I\noZ9n+HAokfb5XfUKm9ij1eIhK07PSX01m3NWnJiT76rptKzEW07AmVmxOicgWZF1ijnx+vkjhBBC\nCCGEEEIIIYQQQgghhBBCCCGEEEIIIYQQQgghhBBCCCGEEEIIIYQQQgghhBBC2KWWU+P1rQbOj6DG\nDUBnK5sSjiRZEWZIToRZkhVhlmRF2ObMuckjMR0Y3MjHJFiwXHFuSVaEGZITYZZkRZglWRG2+a7w\nXQaUoqe2LOHUvOV3AMvRs8a8DXiBq4GDwNfAv4ELjMdeZjwmC9hsXP4FenrHRcAS9FSX04By47GD\nLHpOIjokK8IMyYkwS7IizJKsCNsEObUrYzb621MZkGn8fhh67nHQUzPWeRK4x7j8ClAc9rslQC/j\n8pnh286pKR1/D/zUuJyBnkLT15QnI6JKsiLMkJwIsyQrwizJigXkz+ORqQZ6hl3vCnRBz1UO4EHP\njw7QDZgAtAD86G9wdVwml/c+cNi4/GPgeuAB43oy0A4dQuE8khVhhuREmCVZEWZJViwgG8rWcAGf\nondTnGk6epfDOmAE0C/sdyrschBwG5dTzqhx/IzrxcAXkbUqbCZZEWZIToRZkhVhlmQlAu6z30WY\nsBHIBq40ricClxiX/cAe47ZbORW4Y0B6WI0tQIFxeUgDy1oIjA7CBBU3AAAAqElEQVS73rO+OwpH\nkqwIMyQnwizJijBLshIB2VCOjDrjegAdmKfQB8KvBq4yfvc4+mD2ZcBnYY+ZCYwFVgH5wLPA3eiD\n3jPDlqHOWN6T6CB/AqwHfmvFExJRI1kRZkhOhFmSFWGWZEUIIYQQQgghhBBCCCGEEEIIIYQQQggh\nhBBCCCGEEEIIIYQQQgghhBBCCCGEEEIIIYQQQgghhBBCCCGEEEIIEV/+H+GDbRr94b5dAAAAAElF\nTkSuQmCC\n",
       "text": [
        "<matplotlib.figure.Figure at 0x1077319d0>"
       ]
      }
     ],
     "prompt_number": 5
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "In regression tasks, is very important to normalize data (to avoid that large-valued features weight too much in the final result)"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from sklearn.preprocessing import StandardScaler\n",
      "scalerX = StandardScaler().fit(X_train)\n",
      "scalery = StandardScaler().fit(y_train)\n",
      "\n",
      "X_train = scalerX.transform(X_train)\n",
      "y_train = scalery.transform(y_train)\n",
      "X_test = scalerX.transform(X_test)\n",
      "y_test = scalery.transform(y_test)\n",
      "\n",
      "print np.max(X_train), np.min(X_train), np.mean(X_train), np.max(y_train), np.min(y_train), np.mean(y_train)\n"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "10.2028980046 -4.66702040845 2.40553579807e-15 2.91774920367 -1.93147098641 -1.16236542525e-15\n"
       ]
      }
     ],
     "prompt_number": 6
    },
    {
     "cell_type": "heading",
     "level": 3,
     "metadata": {},
     "source": [
      "A Linear Model"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "Let's start with a lineal model, SGDRegressor, that tries to find the hyperplane that minimizes a certain loss function (typically, the sum of squared distances from each instance to the hyperplane). It uses Stochastic Gradient Descent to find the minimum. "
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "Regression poses an additional problem: how should we evaluate our results? Accuracy is not a good idea, since\n",
      "we are predicting real values, it is almost impossible for us to predict exactly the final value. There are several measures that can be used (you can look at the list of functions under sklearn.metrics module). The most common is the R2 score, or coefficient of determination that measures the proportion of the outcomes variation explained by the model, and is the default score function for regression methods in scikit-learn. This score reaches its maximum value of 1 when the model perfectly predicts all the test target values. "
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from sklearn.cross_validation import *\n",
      "def train_and_evaluate(clf, X_train, y_train):\n",
      "    \n",
      "    clf.fit(X_train, y_train)\n",
      "    \n",
      "    print \"Coefficient of determination on training set:\",clf.score(X_train, y_train)\n",
      "    \n",
      "    # create a k-fold croos validation iterator of k=5 folds\n",
      "    cv = KFold(X_train.shape[0], 5, shuffle=True, random_state=33)\n",
      "    scores = cross_val_score(clf, X_train, y_train, cv=cv)\n",
      "    print \"Average coefficient of determination using 5-fold crossvalidation:\",np.mean(scores)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 7
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from sklearn import linear_model\n",
      "clf_sgd = linear_model.SGDRegressor(loss='squared_loss', penalty=None,  random_state=42)\n",
      "train_and_evaluate(clf_sgd,X_train,y_train)\n",
      "print clf_sgd.coef_"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Coefficient of determination on training set: 0.743303511411\n",
        "Average coefficient of determination using 5-fold crossvalidation: 0.707860049111\n",
        "[-0.07641527  0.06963738 -0.05935062  0.10878438 -0.06356188  0.37260998\n",
        " -0.02912886 -0.20180631  0.08463607 -0.05534634 -0.19521922  0.0653966\n",
        " -0.36990842]\n"
       ]
      }
     ],
     "prompt_number": 8
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "You probably noted the penalty=None parameter when we called the method. The penalization parameter for linear regression methods is introduced to avoid overfitting. It does this by penalizing those hyperplanes having some of their coefficients too large, seeking hyperplanes where each feature contributes more or less the same to the predicted value. This parameter is generally the L2 norm (the squared sums of the coefficients) or the L1 norm (that is the sum of the absolute value of the coefficients). Let's see how our model works if we introduce an L2 or L1 penalty."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "clf_sgd1 = linear_model.SGDRegressor(loss='squared_loss', penalty='l2',  random_state=42)\n",
      "train_and_evaluate(clf_sgd1,X_train,y_train)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Coefficient of determination on training set: 0.743300618503\n",
        "Average coefficient of determination using 5-fold crossvalidation: 0.707861811562\n"
       ]
      }
     ],
     "prompt_number": 9
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "clf_sgd2 = linear_model.SGDRegressor(loss='squared_loss', penalty='l1',  random_state=42)\n",
      "train_and_evaluate(clf_sgd2,X_train,y_train)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Coefficient of determination on training set: 0.743264854792\n",
        "Average coefficient of determination using 5-fold crossvalidation: 0.707816829362\n"
       ]
      }
     ],
     "prompt_number": 10
    },
    {
     "cell_type": "heading",
     "level": 3,
     "metadata": {},
     "source": [
      "Support Vector Machines for regression"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "The regression version of SVM can be used instead to find the hyperplane (note how easy is to change the classification method in scikit-learn!). We will try a linear kernel, a polynomial kernel, and finally, a rbf kernel. For more information on kernels, see http://scikit-learn.org/stable/modules/svm.html#svm-kernels"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from sklearn import svm\n",
      "clf_svr= svm.SVR(kernel='linear')\n",
      "train_and_evaluate(clf_svr,X_train,y_train)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Coefficient of determination on training set: 0.71886923342\n",
        "Average coefficient of determination using 5-fold crossvalidation:"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 0.707838419194\n"
       ]
      }
     ],
     "prompt_number": 11
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "clf_svr_poly= svm.SVR(kernel='poly')\n",
      "train_and_evaluate(clf_svr_poly,X_train,y_train)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Coefficient of determination on training set: 0.904109273301\n",
        "Average coefficient of determination using 5-fold crossvalidation: 0.779288545488\n"
       ]
      }
     ],
     "prompt_number": 12
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "clf_svr_rbf= svm.SVR(kernel='rbf')\n",
      "train_and_evaluate(clf_svr_rbf,X_train,y_train)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Coefficient of determination on training set: 0.900132065979\n",
        "Average coefficient of determination using 5-fold crossvalidation: 0.833648350263\n"
       ]
      }
     ],
     "prompt_number": 13
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "clf_svr_poly2= svm.SVR(kernel='poly',degree=2)\n",
      "train_and_evaluate(clf_svr_poly2,X_train,y_train)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Coefficient of determination on training set: 0.715338649791\n",
        "Average coefficient of determination using 5-fold crossvalidation: 0.609837490871\n"
       ]
      }
     ],
     "prompt_number": 14
    },
    {
     "cell_type": "heading",
     "level": 3,
     "metadata": {},
     "source": [
      "Random Forests for Regression"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "Finally, let's try again Random Forests, in their Extra Trees, and Regression version"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from sklearn import ensemble\n",
      "clf_et=ensemble.ExtraTreesRegressor(n_estimators=10,random_state=42)\n",
      "train_and_evaluate(clf_et,X_train,y_train)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Coefficient of determination on training set: 1.0\n",
        "Average coefficient of determination using 5-fold crossvalidation:"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " 0.850173034292\n"
       ]
      }
     ],
     "prompt_number": 15
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "An interesting side effect of random forest classification, is that you can measure how 'important' each feature is when predicting the final result"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "print sort(zip(clf_et.feature_importances_,boston.feature_names),axis=0)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "[['0.00170257981849' 'AGE']\n",
        " ['0.0197654238108' 'B']\n",
        " ['0.0199455781958' 'CHAS']\n",
        " ['0.0213774533133' 'CRIM']\n",
        " ['0.0222580857806' 'DIS']\n",
        " ['0.0256260086479' 'INDUS']\n",
        " ['0.0267130812242' 'LSTAT']\n",
        " ['0.0267236970887' 'NOX']\n",
        " ['0.0390141196354' 'PTRATIO']\n",
        " ['0.046828306772' 'RAD']\n",
        " ['0.109189407239' 'RM']\n",
        " ['0.286676809021' 'TAX']\n",
        " ['0.354179449452' 'ZN']]\n"
       ]
      }
     ],
     "prompt_number": 16
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "Finally, evaluate our classifiers on the testing set"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from sklearn import metrics\n",
      "def measure_performance(X,y,clf, show_accuracy=True, show_classification_report=True, show_confusion_matrix=True, show_r2_score=False):\n",
      "    y_pred=clf.predict(X)   \n",
      "    if show_accuracy:\n",
      "        print \"Accuracy:{0:.3f}\".format(metrics.accuracy_score(y,y_pred)),\"\\n\"\n",
      "\n",
      "    if show_classification_report:\n",
      "        print \"Classification report\"\n",
      "        print metrics.classification_report(y,y_pred),\"\\n\"\n",
      "        \n",
      "    if show_confusion_matrix:\n",
      "        print \"Confusion matrix\"\n",
      "        print metrics.confusion_matrix(y,y_pred),\"\\n\"\n",
      "        \n",
      "    if show_r2_score:\n",
      "        print \"Coefficient of determination:{0:.3f}\".format(metrics.r2_score(y,y_pred)),\"\\n\"\n",
      "\n",
      "        \n",
      "measure_performance(X_test,y_test,clf_et, show_accuracy=False, show_classification_report=False,show_confusion_matrix=False, show_r2_score=True)\n",
      "\n"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Coefficient of determination:0.849 \n",
        "\n"
       ]
      }
     ],
     "prompt_number": 17
    }
   ],
   "metadata": {}
  }
 ]
}