3.2.3_Support_Vector_Machines_vs_Vanilla_Linear_Classifier.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<p style=\"text-align:center\">\n",
    "    <a href=\"https://skills.network/?utm_medium=Exinfluencer&utm_source=Exinfluencer&utm_content=000026UJ&utm_term=10006555&utm_id=NA-SkillsNetwork-Channel-SkillsNetworkCoursesIBMDeveloperSkillsNetworkCV0101ENCoursera25797139-2022-01-01\" target=\"_blank\">\n",
    "    <img src=\"https://cf-courses-data.s3.us.cloud-object-storage.appdomain.cloud/assets/logos/SN_web_lightmode.png\" width=\"200\" alt=\"Skills Network Logo\"  />\n",
    "    </a>\n",
    "</p>\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Support Vector Machine vs Vanilla Linear Classifier\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h2>Table of Contents</h2>\n",
    "<p>We will be classifying the popular handwritten data set which we can find in the sklearn library and comparing the results of the logistic regression and SVM.  In the Sklearn library, there are several ways to use logistic regression for multiclass applications; in this lab, we will use  the `multinomial` option; this is like Softmax function we discussed before</p>\n",
    "\n",
    "<ul>\n",
    "    <li>Plotting an Image</li>\n",
    "    <li>Preprocess data for Logistic Regression</li>\n",
    "    <li>Logistic Regression with SkLearn</li>\n",
    "    <li>SVM for Image Classification with SkLearn</li>\n",
    "</ul>\n",
    "<p>Estimated Time Needed: <strong>60 min</strong></p>\n",
    "\n",
    "<hr>\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load Important Libraries and Digit Dataset\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/utils/validation.py:37: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.\n",
      "  LARGE_SPARSE_SUPPORTED = LooseVersion(scipy_version) >= '0.14.0'\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/feature_extraction/image.py:167: DeprecationWarning: `np.int` is a deprecated alias for the builtin `int`. To silence this warning, use `int` by itself. Doing this will not modify any behavior and is safe. When replacing `np.int`, you may wish to use e.g. `np.int64` or `np.int32` to specify the precision. If you wish to review your current use, check the release note link for additional information.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  dtype=np.int):\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/linear_model/least_angle.py:35: DeprecationWarning: `np.float` is a deprecated alias for the builtin `float`. To silence this warning, use `float` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.float64` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  eps=np.finfo(np.float).eps,\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/linear_model/least_angle.py:597: DeprecationWarning: `np.float` is a deprecated alias for the builtin `float`. To silence this warning, use `float` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.float64` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  eps=np.finfo(np.float).eps, copy_X=True, fit_path=True,\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/linear_model/least_angle.py:836: DeprecationWarning: `np.float` is a deprecated alias for the builtin `float`. To silence this warning, use `float` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.float64` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  eps=np.finfo(np.float).eps, copy_X=True, fit_path=True,\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/linear_model/least_angle.py:862: DeprecationWarning: `np.float` is a deprecated alias for the builtin `float`. To silence this warning, use `float` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.float64` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  eps=np.finfo(np.float).eps, positive=False):\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/linear_model/least_angle.py:1097: DeprecationWarning: `np.float` is a deprecated alias for the builtin `float`. To silence this warning, use `float` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.float64` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  max_n_alphas=1000, n_jobs=None, eps=np.finfo(np.float).eps,\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/linear_model/least_angle.py:1344: DeprecationWarning: `np.float` is a deprecated alias for the builtin `float`. To silence this warning, use `float` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.float64` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  max_n_alphas=1000, n_jobs=None, eps=np.finfo(np.float).eps,\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/linear_model/least_angle.py:1480: DeprecationWarning: `np.float` is a deprecated alias for the builtin `float`. To silence this warning, use `float` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.float64` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  eps=np.finfo(np.float).eps, copy_X=True, positive=False):\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/linear_model/randomized_l1.py:152: DeprecationWarning: `np.float` is a deprecated alias for the builtin `float`. To silence this warning, use `float` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.float64` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  precompute=False, eps=np.finfo(np.float).eps,\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/linear_model/randomized_l1.py:320: DeprecationWarning: `np.float` is a deprecated alias for the builtin `float`. To silence this warning, use `float` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.float64` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  eps=np.finfo(np.float).eps, random_state=None,\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/linear_model/randomized_l1.py:580: DeprecationWarning: `np.float` is a deprecated alias for the builtin `float`. To silence this warning, use `float` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.float64` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  eps=4 * np.finfo(np.float).eps, n_jobs=None,\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn import datasets, svm, metrics, model_selection\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.metrics import confusion_matrix\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.metrics import accuracy_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/datasets/base.py:550: DeprecationWarning: `np.int` is a deprecated alias for the builtin `int`. To silence this warning, use `int` by itself. Doing this will not modify any behavior and is safe. When replacing `np.int`, you may wish to use e.g. `np.int64` or `np.int32` to specify the precision. If you wish to review your current use, check the release note link for additional information.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  target = data[:, -1].astype(np.int)\n"
     ]
    }
   ],
   "source": [
    "digits = datasets.load_digits()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "target = digits.target\n",
    "flatten_digits = digits.images.reshape((len(digits.images), -1))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Visualize Some Handwritten Images in the Dataset\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1000x400 with 5 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "_, axes = plt.subplots(nrows=1, ncols=5, figsize=(10, 4))\n",
    "for ax, image, label in zip(axes, digits.images, target):\n",
    "    ax.set_axis_off()\n",
    "    ax.imshow(image, cmap=plt.cm.gray_r, interpolation='nearest')\n",
    "    ax.set_title('%i' % label)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Divide Images into Training and Test Set\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "I have set the test size to 20% of the total dataset\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(flatten_digits, target, test_size=0.2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Hand-written classification with Logistic Regression\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Standardize the dataset to put all the features of the variables on the same scale\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "scaler = StandardScaler()\n",
    "X_train_logistic = scaler.fit_transform(X_train)\n",
    "X_test_logistic = scaler.transform(X_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Create the logistic regression and fit the logistic regression and use the <code>l1</code> penalty. Note here that since this is a multiclass problem the Logistic Regression parameter `multi_class` is set to `multinomial`.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "logit = LogisticRegression(C=0.01, penalty='l1', solver='saga', tol=0.1, multi_class='multinomial')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/utils/fixes.py:357: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.\n",
      "  if _joblib.__version__ >= LooseVersion('0.12'):\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "LogisticRegression(C=0.01, class_weight=None, dual=False, fit_intercept=True,\n",
       "          intercept_scaling=1, max_iter=100, multi_class='multinomial',\n",
       "          n_jobs=None, penalty='l1', random_state=None, solver='saga',\n",
       "          tol=0.1, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "logit.fit(X_train_logistic, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred_logistic = logit.predict(X_test_logistic)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Get the accuracy of the logistic regression\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.825\n"
     ]
    }
   ],
   "source": [
    "print(\"Accuracy: \"+str(logit.score(X_test_logistic, y_test)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Lets plot out the confusion matrix, each row of the matrix represents the instances in a predicted class, while each column represents the instances in an actual class.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "label_names = [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]\n",
    "cmx = confusion_matrix(y_test, y_pred_logistic, labels=label_names)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Accuracy is fine and above 80% but we can see some heavily misclassified values, The classifier had a hard time classifying <code>8</code>\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAhUAAAHBCAYAAADeo0naAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/NK7nSAAAACXBIWXMAAA9hAAAPYQGoP6dpAACiRklEQVR4nOzdd3yN5//H8VeSk53IkASxV1JEELFHrZpfVavqq/iitLVrj/5UqVmqNbusLqVGqT1a2qq9gtgzQhCRvZPz+yNyOM0+8058nt/HeXybc693rvs+x5Xrvu7rslCr1WqEEEIIIfRkae4AQgghhCgapFIhhBBCCIOQSoUQQgghDEIqFUIIIYQwCKlUCCGEEMIgpFIhhBBCCIOQSoUQQgghDEIqFUIIIYQwCKlUCCGEEMIgpFKhIOnp6axZs4bOnTvj7++Pr68vvr6+ZsmSeex79+6Z5fgvu1atWuHr68uxY8dMdswtW7bQo0cPateuLedf6MUc169QBpW5AxjbnTt32LBhA8eOHSM0NJTo6GgcHBwoW7YsgYGB/Oc//8Hf39/cMQFYtmwZS5cuxcLCgipVquDs7GzuSIXCixWvnj178sknn+S4blxcHE2bNiU+Ph6APn36MG3aNIPk2Lx5M6GhobRp04Zq1aoZZJ+msnnzZiZPngxAhQoVcHd3B8DW1tacsTSePn3Kjz/+yJ9//smtW7eIj4+nWLFiFC9enCpVqlCvXj1atmyJt7c3arWaNm3acO/ePYYMGcLYsWPz3H9ERATNmzcnJSWFJUuW0LZtW0A511ZRcezYMY4fP061atVo06aNueMIIyiylYq0tDQWLFjA2rVrSUtLw8LCgjJlylC6dGliY2O5du0aFy9eZO3atbRu3Zrly5ebNa9arebHH38E4LPPPqNjx45mzVOxYkUArK2tzZqjoHbt2sXUqVOxt7fPcXnml76hbdmyhePHj1O6dGm9KxVly5bFxsYmx9/D0H744QcAxo0bx+DBg01yzPwKCgri3XffJSIiAgBPT0/Kli1LWload+/e5dq1a+zatYuoqCiGDh2KhYUFXbt2ZcmSJfz666+MHj0aKyurXI+xbds2UlJScHNzo0WLFtmuY85rq6g4fvw4S5cupWvXrlKpKKKKZKVCrVYzatQo9u3bh42NDUOHDqV3794UL15cs058fDwHDx5kxYoVHD9+3IxpM0RERPD06VMAWrZsaeY0sHv3bnNHKLDKlStz48YN9u7dS5cuXbJdZ9OmTVrrKtXatWtNerzMslDCtfei+Ph4hg0bRkREBLVq1eLDDz/UallMT0/nwoUL7NixAxcXF8373bp1Y9myZTx69IjDhw/TvHnzXI+zefNmAF5//XVsbGyyLC9K15YQxlQk+1SsWrWKffv2YW1tzcqVKxk+fLhWhQLAwcGBjh078uuvvzJkyBAzJX0uMTFR89+m+uu0qOnatSvw/B+If7t9+zanT5+mSpUqirnlpRSZ15+dnZ2Zk2g7ePAgjx49wsrKiqVLl2Y5b5aWlvj7+zN58mT69Omjed/b25uGDRsCGS1Iublw4QJXrlwBMioj2ZFrS4j8KXKVivj4eL7++msAhgwZQv369XNd38rKKttKhVqtZvv27QwYMIAGDRrg5+dH8+bNGTt2LMHBwdnua/Pmzfj6+tK3b1/Nzz179qROnToEBATQt29fDh8+rLXNvXv38PX1pVWrVpr3MjvJ+fr6smTJkmz3nZ2+ffvi6+ub7Rff8ePHGTFiBM2aNcPPz4/AwEDatWvHyJEjNX9hvSi3jnqmKpuCatCgAaVLl9b0n/m3zN8zp384MgUHB/PFF1/w1ltv0bx5c/z8/GjQoAH9+vXj119/Ra1Wa61/7NgxfH19NS1ekydP1jqHL56zSZMmac5rdHQ08+bNo23bttSsWVPrL+DsOrpdvHgRPz8/fH19OXDgQLbZR48eja+vLz179iQlJSWPEiNLZ+DWrVtr3ps0aZLWug8ePGDmzJm0a9cOf39/6tatS48ePVi9ejVJSUnZ7v/F3+P69euMGzeOZs2aUb16dWbNmpVnvpCQEADc3Nzw8vLKc/0Xde/eHYADBw4QHR2d43qZlY4aNWrwyiuvZLuOoa6t3OT32oCM743Mc1GrVi3q1KlDt27dWLVqVY7nIiQkhGnTptG2bVv8/f2pXbs2rVq14n//+x9ffvlllls3eXXWzc930r/3t3TpUiCjzF/8jPy7Q3pBv6+EchS5SsWhQ4eIjIzEysqKt99+W6d9pKamMmrUKMaOHcs///yDnZ0dvr6+xMfHs337dnr06MEvv/yS6z6mTJnC5MmTefz4MRUqVECtVnP8+HHeeecd9u/fr1nP1taWgIAA/Pz8NO8FBARoXqVKldLpd3jRpk2b6NevH3v37iUxMZHKlSvj7e3N06dP2bNnD1988UW+92XKsimozHvparWaX3/9VWtZWloav/76KyqVKsfm60wffvghy5cv5/r16zg6OuLr64uNjQ3Hjh1j4sSJTJgwQWt9Z2dnAgICcHJyAjI6Or54Dn18fLIc4+nTp3Tr1o3Vq1djbW1NlSpV8uwUWaNGDcaPHw9klGFYWJjW8l9++YVdu3bh5OTEZ599lq/+MJkZM/n5+Wneq1Chgub9EydO0LlzZ3744QdCQ0OpXLkynp6enD9/nrlz59K7d2/N7bvsnD59mu7du7N37148PT2pWLEiFhYWeebLLNPw8HDu3LmT5/oveu211yhWrBhJSUns2LEj23WSk5PZvn078LwSkh1DXVv5kde1sW/fPjp16sQPP/zAgwcPKFOmDB4eHly6dIl58+bRt29fYmNjtfZ5+fJlunbtyvr16wkLC6Ns2bJUqlSJ5ORkjh07xqJFi3j8+LHe2XPz4vdZ8eLFtT4jL16Dhvy+EmagLmJmzpyp9vHxUXfu3FnnfSxZskTt4+OjrlWrlnrv3r2a95OSktSzZ89W+/j4qKtXr64OCgrS2m7Tpk1qHx8fdY0aNdT169dX//3335plcXFx6vfff1/t4+OjbtmypTo9PV1r25CQELWPj4/ax8cn20yZ+3777bdzzP3222+rfXx81Js2bdK8l5qaqq5fv77ax8dH/d1336lTUlK0trl+/bp67dq1WfaVmSUkJMTsZZOXzKxBQUHqkJAQta+vr7p169Za+zl48KDax8dH/e6776rVarV64sSJah8fH/XHH3+cZX/btm1TX7lyJcv7586dU7dt21bt4+Oj3rlzZ5bl2ZX/v2Uet1q1auquXbuq79y5o1mWkJCg+e+WLVuqfXx81EePHs2yj3fffVft4+Oj7tOnjzo1NVWtVmecx1q1aql9fHzU27Zty/H4OcnpfKvVavWTJ0/UDRs21JRfRESEZtmFCxfUr776qtrHx0f9/vvvZ9k28/eoVq2aevz48eqYmJhsf9+c3L17V12tWjXNtfHzzz+r79+/n+/f66OPPlL7+Pioe/bsme3ynTt3qn18fNQ1a9ZUR0ZGZllu6GsrN/m5NoKDg9V+fn5qX19f9fLly7XK8Pbt2+ru3burfXx81BMnTtTa99ChQ9U+Pj7qcePGaZ0DtTrj/P7www/q8PDwbH/37K4JtTr376Scrt/Fixdnmy+Trt9XQjmKXEvFw4cPgYze87qIj49n9erVAIwYMYLXXntNs8zGxobJkycTGBhIamoqX375Zbb7SElJYcqUKTRp0kTznoODA9OnT8fa2prQ0FDNPVxji4iIIDIykmLFitG3b19UKu2+uZUrV6Zfv3752ldhKJsyZcpQv359QkJCOHnypOb9zCbT3P4azdS5c+dsWxf8/f356KOPgJzvreeXpaUlS5cupVy5cpr38tufYfbs2Xh5eXHixAmWL19OUlISH3zwAQkJCXTr1o3OnTvrle3f1q1bR0REBO7u7ixatAg3NzfNsho1ajBnzhwg4zbD5cuXs91HpUqVmDNnjqblAfL3+5YtW5bJkydjaWlJaGgo06ZNo0WLFjRp0oTBgwfz9ddfc/fu3Ry3zzzf586dy7bzZOatjzZt2mh19MyOIa6t/Mjt2vjss89ITk5m8ODBvP/++1plWL58eZYsWYKDgwPbtm3TfBfC8464gwYN0joHAO7u7vTp0ydLvzNzMOT3lTCPIlepyGz2c3Bw0Gn7kydPEhsbi729Pb179852nYEDBwJw+PBhUlNTsyx3dnbm9ddfz/K+l5cXpUuXBsj1i9CQ3N3dsbW1JTo6Osf78PlVWMom88s98x/+yMhIfv/9d9zd3XN8XPDfQkND+frrrxk9ejT9+/end+/e9O7dm4ULFwLk2Hckvxo3boy3t7dO27q7u/Ppp59iaWnJihUrGDp0KFeuXKFChQr83//9n165snPo0CEA/vvf/2bbibhRo0ZUr14dyOhYmZ033ngjz8c6c9K3b182bNhAp06dcHR0BDJuh/z5558sXLiQtm3bMnny5Gwf56xZs6amgvjviuCjR4/4+++/gfxXCAxxbeUlp2sjNjZW0++oV69e2W5bqlQp/Pz8SEtL03qqLXN/27ZtIy0tzSA5jcGQ31fCPIpcpSLzS0fX58Vv3boFZPyFlFPFJPNLKiEhgQcPHmRZXr58+RzvF3t4eOiVr6CsrKwYMGAAAEOHDqVTp07MmjWL3377TesvmfwoLGXTrl07nJyc2L17N/Hx8ZoxCF5//fV89TP4/vvvadeuHQsXLmTXrl0cPXqU06dPc/r0aS5cuABk/GOij8qVK+u1fcOGDXn33XdJS0vj77//xtramkWLFulcmc5N5nnPrvUmU+aymzdvZru8SpUqemWoWbMmn332GSdOnGDr1q3Mnj2bLl264OrqilqtZvPmzXzwwQfZbptZEfj3P6i//voraWlpeHt706hRo3zl0Pfayo+cro2rV69qxtwZP368pqL779f169cBtPrcDBw4EAsLC1auXMmrr77KpEmTWLduHdeuXTNIZkMx5PeVMI8iN05FyZIlgee9xgsqLi4OeP4PXHZe7IWeuf6Lcvtit7TMqMelp6frlE8Xo0ePxtvbmx9//JErV65ovnQsLCxo1KgREydOzLHX+4sKS9nY2dnRsWNHNmzYwJ49ezR/VeanZ/6ZM2c0oyb26dOHN954gwoVKuDo6IiVlRUhISG0adMm21aYgjDEP/6NGzdmxYoVQMatmczWAkPLz3n39PTUWvffDPWYtJWVFa+88gqvvPIK3bt3JyYmhsmTJ7Nv3z4OHjzI2bNnqV27ttY2r7/+OgsWLNC0TLz66qsAmg6Xb7zxhubay4s+11Z+5XRtZD7BolarOX36dJ77efEx9aZNm7JmzRq++uorTpw4wZYtWzS3fipXrswHH3ygdTvTnAz1fSXMo8i1VNStWxeA69eva0bgK4gXm1dz8ujRoyzrG1vmX/fqfz3O+KKc/sK3sLCgV69ebNu2jSNHjrB06VL69u2Lu7s7//zzD/3798/XXwFKLZvsZH7JL1myhEuXLlGjRo18zaOS+UXbvn17pk2bhr+/P8WKFdM03ef2hIMpRUdHM3HiRCCjMnbq1CnNiKyGlp/znvnkgKnPubOzM3PmzNFUCs6ePZtlHXd3d82gXpnn9+zZs9y4cQMLC4sCVwh0vbb0lVnZKFasGFeuXMnzNWLECK3tGzZsyOrVqzlx4gRr1qxh+PDhVK1alRs3bjBixAj++uuvbI+b03dOQkKCYX/BZwz1fSXMo8hVKpo3b46rqytpaWmaoYcLolKlSkBGS0dOH5rMJkN7e3uDPPKZH5l/6eX2xZ6fR+7c3d157bXX+PDDD9m7dy+lS5cmMjKSXbt25bmtUssmO3Xq1KFSpUqaMQXye888c/3AwMBsl585c8YwAfX04Ycfcv/+ferUqaPp5zFv3jyjdADOPO+5NZVnLtP3to4unJ2dNXOV5DQ2R2ZF4MCBA0RFRWlaGOrVq1fgTt26Xlv6qlKlChYWFkRHR2v+eteFvb09jRo1YsSIEfz222+0a9cOtVrNunXrtNbLrMQ8efIk2/1k3hYriPw8Rvwifb6vhHkUuUqFo6Mj77zzDgBff/11nkNwp6WlaQbLgoyWDicnJxISEli/fn2222Q+AdG0adMsvZONJXPMgHv37mVbsdi2bRsxMTEF2qeTk5OmGfHFFoacKLVscjJ48GAaNWpE48aN6dSpU762yexNn115JCYm5toakFnxe7HZ2Rh+/vln9uzZg7OzMwsWLKBjx4706tWLpKQkxowZY/DjZ94uWLduXbb7Pnr0KBcvXtRa11AiIiLyvB128+ZNzT98mXPW/Fvz5s3x9PQkOTmZzZs3s3PnTkD3CoEu15a+3N3dNaOEGmquIgsLC80YEf++5suXLw9k3/oTExOT49gfucn8fOlyjRb0+0qYR5GrVAC88847tGrVipSUFAYNGsTSpUuz1LYTEhLYs2cP3bp106pUODg4aDoKffHFF1o9kJOTk5k/fz7Hjx9HpVLx7rvvmuYXIqMjXOnSpUlJSWHGjBlaLQVHjhxh9uzZ2XYUu379OlOmTOHkyZNZvpwPHz7MkSNHgIyOcHlRatnkpFu3bqxZs4bVq1fj6uqar23q1asHwE8//aT1ZfrkyRNGjBiRZcCpF2U+Anj8+PFcb1Pp49q1a5pHOGfOnEmZMmWAjMGwqlatyvXr15k9e7ZBj/nWW2/h7u5OeHg4Y8eO1eqkGhwczJQpU4CMxzINfa97586ddOrUibVr12Ype7VazV9//cXQoUNRq9WUKlWKpk2bZrsfKysr3njjDQA+//xzYmJicHJyol27djrl0uXaMoTx48dja2vLjh07mDp1apYBq5KTk/nrr78YOXKk1vsjR45k7969WVoY7969y4YNG4Cs3wGZo/x+++23Wo8KP378mDFjxhT4jxh4XlEJCgrK9natIb+vhHkUuY6akFH7XrJkCfPmzePHH39kyZIlLF26lLJly+Lq6kpcXBwhISEkJycDaKY5zvTee+9x9epV9uzZw9ChQ/H29qZ48eLcvn2bmJgYrKysmD59ukkvbEtLS6ZMmcKIESPYs2cPf/31FxUrVuTp06fcv3+f7t27ExISkqVlJiUlhU2bNrFp0ybs7e0pV64cNjY2PHz4UFPbb9OmDe3bt89XDiWWjSH17NmTDRs2cOPGDd566y3Kly+Pg4MD165dw8LCgmnTpvHhhx9mu+3rr7/Ojz/+yK5duzhz5gylS5fWdCycOnWq3tlebIno2bMnHTp00Cyzs7Pjs88+o0ePHqxfv57GjRvn+5zmxd3dnS+++IL333+f/fv38+eff1K1alUSEhI0T3vUqFEj12nBdWVhYcHNmzeZPXs2s2fPxtPTEy8vL1JTUwkLCyMqKgrI6ES6dOnSXMe+6NatG998843mr+SOHTsWunl2atSowZIlSxg7diwbN25k8+bNVKhQgWLFihETE8Pdu3ezvQX0zz//sGfPHlQqFWXLlqVYsWJERUVx584d1Go1FSpUYPjw4VrbDBgwgG3bthESEkLXrl0pX748tra2XL9+HS8vL4YNG8bnn39eoPxNmjTBw8OD0NBQXn31VSpWrKgZLfT77783+PeVML0iWakAUKlUTJ06lf/+97/88ssvHD16lNDQUO7fv4+dnR1VqlShbt26vP7661kmAFKpVHzxxRds376djRs3cunSJR4/foybmxuvvvoqAwcOpEaNGib/ndq0acO3337Ll19+yYULF7h58yaVKlVi6NCh9OzZM9sx+CtUqMCsWbM4cuQIwcHBhIWFERcXh7OzM40bN6ZLly68/vrr+b7XqdSyMRRHR0d+/PFHTUtMaGgorq6utGnThvfeey/LwEEv8vf3Z9myZaxevZrLly9z5swZgz7lM3v2bK5evUrlypWzraT4+PgwadIkPv74Y/7v//6PmjVrasb+0Ff9+vX57bffWLlyJX/++SfXrl1DpVLh5+dHp06d6NOnT57DjOuiV69e+Pr6cvjwYU6ePElYWBi3bt0iJSUFFxcXGjRoQIsWLejZsyfOzs657qtSpUrUqVNH0y/GVH0hDO3VV19l165d/PDDD/z555/cvXuXe/fu4enpSa1atWjYsGGWacXnzZvH4cOHOXPmDA8fPiQkJAQ7Ozv8/Pxo06YNffv2zdLJ1tnZmXXr1vHFF19w8OBB7t27h4eHB7169WLEiBH88ccfBc7u4ODAmjVrWLJkCadPn+bixYtaT1IZ+vtKmJ6F2ljttEIIIYQwC7VaTf/+/TWTEu7cuTPbjtQhISEsXryYI0eOEBUVRcmSJWnfvj3vv/++To++F8k+FUIIIcTLbP369Rw7dizXVp3g4GC6dOnCtm3b8PLyonXr1pqHF956660sE9Plh1QqhBBCiCIkLCyMTz/9lGbNmuU4HUBaWhpjx44lLi6OsWPHsnnzZj7//HN2795Ny5YtuXLlCp9++mmBjy2VCiGEEKIImTZtGunp6UyfPj3HdX7//Xdu3ryJj48PgwcP1rxvY2PDjBkzUKlUbNq0qcAD/kmlQgghhCgifv31Vw4dOsSoUaM0j5xn5/fffwcy5rP59y0SLy8v6tatS0pKimZCwfySSoUQQghRBISHhzNnzhz8/PyyfRrwRZljj/j5+WW7PPMpvoKO0iuVCiGEEKIImDFjBrGxscyaNUszX1FO7t+/DzyfhPPfSpQoATyfuiC/iuw4FUIIIURh0bp161yXvziCcXb27t3Lnj17GDJkSIFmnc5pALjMcUtymnk4J4WqUpFw4Ou8VzIj5w4fmzuCEEKIZ1KTC/ZXti5Swm8a/Rh5iYyM5OOPP6ZcuXIMGzasQNvm9MiprkNYFapKhRBCCFEU5dUSkZs5c+YQHh7OmjVrch2q/kUODg5ERUVlOwcLPJ/a/t8jreZFKhVCCCGErtLTzJ2AAwcOYGtry/Lly7PMYJs56dzEiROxt7enT58+tG/fHm9vb6KioggLC8v2dknmBH45jXORE6lUCCGEELpSG25+H30kJSVlmVDyRefPnwee992oVq0aly5d4uLFi7Ro0SLL+sHBwQAFnnlYKhVCCCFEIXby5Mkcl7Vq1YrQ0NAsc3+0bNmSzZs3s2fPniz9MB4/fsypU6dQqVQ0b968QFnkkVIhhBBCV+nphnmZWKtWrahQoQJXrlzhm2++0byfkpLCtGnTSElJoXv37ri7uxdov9JSIYQQQuhIrZDbHwWlUqlYuHAhffv2ZcGCBezevZty5cpx7tw5QkND8fHxYfz48QXer7RUCCGEELoqpC0VkDGa5q+//krnzp0JCwtj3759WFpa8s477/Dzzz/j7Oxc4H1aqHV9GNUM8jNOxe9nr3E4+DbBd8J4HBVHZFwCdjbWVCrpTru6r/Bm81pYq7RHGlux/R++2nkk1/1umfY/KpYsnus6hhinokP7VoweNYQ6dfywtbXlytUbrF27nuUr1uj83LAhSb6imU3yFd1sL3M+U4xTkXzvvEH2Y1OmpkH2Y25FrlLxv4XrOHvjPjYqKzxdnHBxtCM8Oo5HkRnzwlcvV4IvR/agmMPzZ3kzKxUl3Zwp6Z59zWzOgE6Uci+W67H1rVRMGD+M2bOmAHDjxm1i4+Lxq+GLlZUV237bQ/ceg8z6BSD5imY2yVd0s73s+UxSqQg5Z5D92JStZZD9mFuRq1RsPXIB7+Iu1K7sjfULY58H3brP+G9+42FkLG82r8WUt9polmVWKt7t2Ij3/9NY53z6VCoaNqjLn4d+Ra1W0+9/I1i/fisA/v7V2bn9R0qW9GLCxBl8tugrnY+hD8lXNLNJvqKbTfKZqFJx57RB9mNTPsAg+zG3ItenoksjP+r5lNWqUAD4V/RmbI8WAPxx7oYZkuVuyuRRWFpasnLVOs0HCyAoKJhxEzIqKxPGD0elMk/fWslXNLNJvqKbTfIJcyhylYrcVCyR8WhMYnKKmZNoc3Z2onXrpgCsXr0uy/KNG7cTFRWNh4c7LVvo3pKiK8lXNLOB5Cuq2UDymYw63TCvIkLvSkVCQgLh4eGEh4drxgpXqnO3HgBQrVyJbJefvBrCuG9+Y/DnGxj3zTZW7z1OeFTBZmjTRZ3aGZ2TEhISOH0ma6ef1NRUTp7MuG9Xv77pm8gkX9HMBpKvqGYDyWcyhfjpD2MocJtSQkICmzZt4sCBA1y+fJnIyEit5S4uLrzyyiu0adOGbt264eDgYKisOklLTyc8Ko6DQTdYvPUv7G2tGdGlabbrnrp+T+vn/Weu8dWOI0x+qzVdGvkZLWOVKhUBuBtyn7S07MeRv3nrDq1pplnXlCSf7pScDSSfPpScDSSfMI8CVSqOHDnCuHHjiIiIyLE3bmRkJEePHuXYsWOsWLGC+fPn06RJE4OELYgffj/Fgo0Htd5rWasKwzo3oYq3h9b7ni6ODGrXgFa1q1DGwwVbaxWXQx7xze5jHL54i+k/7MHV0Z5X/StjDG5uLgBEPo3KcZ3IyCitdU1J8ulOydlePKbkKzglZ3vxmJLPuArr4FfGku9KxaVLlxgyZAgpKSm8+uqrdOjQgRo1alCyZElNa0RCQgJhYWFcvHiRXbt2cfDgQd577z02bNhAtWrVjPZLZMfL1Ynalb1JTUvnQUQ0T6LjOXE1hF0nLjO0c2OsLJ/f+enRLOujPLUrl2bp0K6M/Xobv5+7zoJNB2les1KOc8/rI3Oq2uTk5BzXSUrKWGafz2ltDUny6U7J2UDy6UPJ2UDymUwRunVhCPmuVCxfvpzU1FQ+++wzOnbsmO06Tk5OVKlShSpVqtClSxd27tzJmDFjWL58OUuWLDFY6PxoG+BL2wBfzc/nbz1g5rp9rNxzjOj4RKb2bpPL1hksLCwY9UYzfj93nZDHkVwNfYxvGS+DZ01MTATAxsYmx3VsbTOWJTxb15Qkn+6UnA0knz6UnA0knzCPfHfUPHnyJHXr1s2xQpGdjh07EhgYmOsMaqZSs2Iplg7tho3Kik1/B3H/SXS+titfwh0Xx4xacsjjSKNke/qs+c81lyY+V1cXrXVNSfLpTsnZXjym5Cs4JWd78ZiSz8jk6Q8t+a5UxMfH4+npWeADeHh4EB8fX+DtjMHL1QnfMl6kq9VcDX2U7+1UVhnFlJZmnBN//fotAMqV9cbqX+NrZKpUsbzWuqYk+XSn5GwvHlPyFZySs714TMlnZOlphnkVEfmuVJQtW5YTJ04UqIIQFxfHiRMnKFu2rE7hjCHt2f2vtLT8DST6NDaeiJiM39nLreCTq+THmbMXSE5Oxt7enoA6Wcd/V6lUBAZm9Ps4ftwwo7cVhOQrmtlA8hXVbCD5TEZaKrTku1Lx+uuvEx4ezoABAwgODs5z/eDgYAYNGkRERASvv/66XiENJfRJFFfvPQbAp0z+Wl2+P3AKtRqc7G3xK1/SKLliYmI5cOAvAAYM6J1leY8e/8HFpRjh4REcPJT7xGfGIPmKZjaQfEU1G0g+YR75rlQMHDiQBg0acO7cObp3785rr73G6NGjmTt3Ll988QWLFy9m7ty5jB49mrZt29K9e3fOnj1L/fr1GThwoDF/B43guw9Zvv0w98Ijsyw7fPEWw5duJjU9naY1KlLW0xWA6/fDmbVuP9fvh2utn5SSyre7j7Fm7wkABrxWL8vspoY0Z+5i0tPTGTSwN716ddG87+9fnQXzPwJgwcLlpKSYZzRQyVc0s0m+optN8pmIDH6lpUATiqWmpvLNN9+wdu1arUGvMh+zfHFXLi4u9O/fn8GDB2NtbW2QsHlNKHbiagiDP98AgEcxR7xcnTSPlMYkJAFQo3xJlg7riptTxmOwl0Me8dac7wFwc7LXzER6M+wJicmpAHRt7Me0Pm3zfJxU31lKJ08aycwZE4Gss/Xt2LGfrt0HkG7Gi0/yFc1skq/oZnvZ85liQrGkC/sMsh9bv9cMsh9z02mW0rS0NE6fPs3ly5e5f/8+cXFxqNVqHB0d8fb2plq1agQEBOTY+UZXeVUqouMT2X4smONX7nLjwROeRMeRkpaOq6MdPmW8aBvgQ6f61TUdLzO3WX/oLGdv3ud2WARPY+NJSUvH3cmemhW96dakJo2rV8hXPn0rFQCdOrZh1MjBBATUxNrammvXb7F27XqWLV9t1g++5Cva2SRf0c32MueTSoXpFbmpz83JEJUKIYQQhmGSSkXQHoPsx9a/nUH2Y24yn6wQQgihI7W66DwOaggv1dTnQgghhDAeaakQQgghdFWExpgwBKlUCCGEELpSQEdXJZHbH0IIIYQwCGmpEEIIIXQltz+0SKVCCCGE0FURmgzMEKRSIYQQQuhKWiq0SJ8KIYQQQhiEtFQIIYQQupKnP7RIpUIIIYTQldz+0FKoKhUenWebO0Kuohd2yXslM/KcuNPcEXKUlKrgqY0LAVuVYWYCfhnJtac7byd3c0cQClOoKhVCCCGEosjtDy1SqRBCCCF0JZUKLfL0hxBCCCEMQloqhBBCCB3J1OfapFIhhBBC6Epuf2iR2x9CCCGEMAhpqRBCCCF0JeNUaHlpKhXly5ehZaumBAbWIrBubapVr4pKpeLjjxcwf95Sk2T44+ZjDt8JJ/hRDI/jkohMTMFOZUUld0faVfWip18ZrK20G4/C45I4EhLBxYfRXHgYzdXwGFLS1bxR3ZuPWlUzSW5QRvnlR4f2rRg9agh16vhha2vLlas3WLt2PctXrEGtVku2bCj53Co524uUem6Vnq9tx1a0bNMU/9o1KFHSE1d3VxISErh25Sbbt+zm+1XrSUlJNVu+fJHbH1pemkrFsGEDGTZ8oFkzfHfmDmcfRGFjZYmnow0+Hk6ExyUTFBZFUFgU2y+H8dUbdXC2fT6Q0Z5rD1nw9zUzps6ghPLLy4Txw5g9awoAN27cJjYunlr+1fni809o3boZ3XsMMtsXqJKzKfncKjlbJiWfW6Xne3d4f+o1DCAxMYlHYY+5dOEKXiU9Caxfm8D6ten2Zmf+23Uw0dExZsmXL9JSoeWl6VMR/iSCnTv3M2PGQt7o0p8tW0w/umTX6t5880YAfw95le39mvDjm/XZM6Apa3sEUsLJlkuPY1h69IbWNo42KhqWdeedwAos6uTPW/5lTJ4blFF+uWnYoC6fzJxEWloaffoOxbdaE+oGvka9Bu0JC3vE653b8cHoIZItG0o+t0rOBso/t0rPt+77zbzZeQDVyzWkaUAHOrfpTQO/NnRp24f7oWH416nB+A9Hmi2fKLiXplIxf95SevZ4h3lzl7Bv3yHiYuNNnuH1at4ElnHLcovDv6QLY5tWBeCPm+Fay96o7s2KLnUY1rAyLSp64mJrnuGYlVB+uZkyeRSWlpasXLWO9eu3at4PCgpm3ISPAZgwfjgqlekb55ScDZR9bpWcDZR/bpWeb+O6rRw9fJLUVO1bHGdOBjHzw08BaNexpTmi5V96umFeRcRLU6lQugqujgAkpsozzwXl7OxE69ZNAVi9el2W5Rs3bicqKhoPD3datmgs2YRBKP3cKj1fXq5fvQWAvYO9mZPkQZ1umFcRIZUKhQgKiwKgmqezmZMUPnVqZ3Q+S0hI4PSZ81mWp6amcvLkOQDq1w+QbMIglH5ulZ4vL3Xr1wLgwrlgMycRBSGVCjNKS1fzMDaRDefvsejwNeytrRjRqLK5YxU6VapUBOBuyH3S0rJv6bl5647Wuqai5GxCP0o/t0rPlx1LS0tKepeg78BeTJ0xlrjYeObO/MLcsXIntz+0mORG2vLly7l37x6zZyt76nJT+fHs3SxPdLSs5MnQBpWoUtzJTKkKLzc3FwAin0bluE5kZJTWuqai5GxCP0o/t0rP96JB773NR7Mnar23e/sBFsxZytVL182UKp+KUIXAEEzSUnHo0CG2bNliikMVCl5OttQu5YJfiWIUd7AB4MS9p+y++pC0dPM/017Y2NnZAZCcnJzjOklJGcvsn61rKkrOJvSj9HOr9HwvCnvwiBNHT3PmVBCPHmZ0Vm/UrB5dunXA0lIa1AuTl2acCiV5rUoJXqtSQvPz+bAoPjl4mZWnbhOVlMLUFq+YMV3hk5iYCICNjU2O69jaZixLeLauqSg5m9CP0s+t0vO9aMfWvezYulfzc+26NZnz2TRGjB2Cq5sLU8d9YsZ0eShCnSwNoUCVivv37+t0kNxqygJqlnRhyX9q0/n7f9h8MZQBAeXxLqbwHs8K8vRZ865rLk24rq4uWuuaipKzCf0o/dwqPV9uzp46z/96DeWv07v4b/8eLP98JaH3Hpg7Vvbk9oeWAlUqWrVqhYWFRYEPolarddruZeLlZIuvhxPnH0Zz9UmsVCoK4Pr1jEfPypX1xsrKKttOaZUqltdaV7IJfSn93Co9X14ehj0m+MJlAgJrUd3PV7mVCqFFp9sfxYsXL9D6kZGROfY+Fs+lPutPIf0qCubM2QskJydjb29PQJ2anDh5Vmu5SqUiMDDj8bTjx09LNmEQSj+3Ss+XHyqrjH+irFRWZk6SC7n9oaVAlQpvb28ePHjA5s2b8fLyyvd2vXr1IigoqMDhXib3oxO49iQWAB8PeQKkIGJiYjlw4C86dGjNgAG9s3x59ujxH1xcihEeHsHBQ0ckmzAIpZ9bpefLS5my3lTz8wHg0oUrZk6TC7n9oaVA3Wpr1qwJwMWLF40SpigLfhTNimM3uReVkGXZ4TtPGP7bWVLT1TQtX5yyLg5mSFi4zZm7mPT0dAYN7E2vXl007/v7V2fB/I8AWLBwOSkpKZJNGIzSz62S89WsVZ0xk4ZSrnzW+Yxebd2Etb+swNramgN7/+TO7Xsmz5dvMqKmFgt1Aaan+/bbb1mwYAFDhw5l5Mj8T/Ly5ptvcv78eS5duqRTyEyODhV03rZhw7qs3/CN5mcnJ0fs7GyJi4snIeF5z+fGjToRGqrbvbvH8zrmuOzkvacM/jWjidHDwQYvJ1tS0tSExSYSk5Qx7n0Nr2Is6VwLN/vnvbXDYhLpvf645ufE1DQSU9OxsbLEwfp5k+CiTv7ULuWaaz7PibpPxmTs8ktK1f9LbfKkkcyckfGse+ZsjH41fLGysmLHjv107T6AdDP9VWHsbLYq3eeEMcVnQ8nZ9L32lHzdGTuft5O7zrkaNglkw2+rAXgU9pgH9x9ibWNN6TKlcHEtBmR02OzfayhPIyJ1OsbdiKwjiRpawmbDjL9k322KQfZjbgW6/dG4cWNat26Ng0PB/pIeNmwYERERBdrG0KytrfHwyPoBcHR0wNHx+e9jZWWce3c+Hk6Mb+bD8XsR3IyI4/bTeFLS0nGxs8a/fHHaVvGio29JVP96JjtdrSYyMeuXXnJaOslpz78IUtOM2w/D3OWXH3PmLiYoKJhRIwcTEFCTkiW9OH/hMmvXrmfZ8tVm/WJXcjYln1slZ8uk5HOr5HyXLlzlo0lzaNK8IT6vVKZy1YpY21gTGRHJqRPn2P7rHrZs2K78/nhy+0NLgVoqzE2flgpTyK2lQgn0aakwNkO0VLzM9GmpeNnJtac7fVoqTMEkLRUbZhhkP/ZvTjPIfsxNhioTQgghhEHIiJpCCCGErgpPY79JSKVCCCGE0JX0qdAilQohhBCikFu/fj1HjhzhypUrPHnyhLi4OFxcXKhZsyZvvfUWLVu21Fo/PDycgwcPcujQIa5du8aDBw+wsrKiatWqdO7cmbfeeguVquBVBKlUCCGEELpSSEvF6tWrCQkJwcfHh4CAAOzs7AgJCeHgwYMcPHiQgQMHMnHi8+nl586dy2+//YZKpaJGjRpUq1aN8PBwzp49y9mzZ9m9ezfffPMN9vYFmzJCKhVCCCGErhQycNWcOXPw8fHB0dFR6/2TJ08yePBgVq1aRfv27alVK2NodldXV8aOHUuPHj1wd3/+FM+tW7cYOHAgJ06c4Msvv+SDDz4oUA55+kMIIYQo5OrUqZOlQgEQGBhIhw4dADhy5Plw7B9++CFDhgzRqlAAVKxYkbFjxwLw22+/FTiHtFQIIYQQulLI7Y/cZPaNsLGxyWPNDK+88goAjx49KvixCryFEEIIITIo/JHSS5cusWvXLqysrGjWrFm+trlz5w4Anp6eBT6eVCqEEEIIXSmspWLTpk2cOHGClJQUQkNDOXv2LCqViunTp1O1atV87WPNmjUAtGnTpsDHl0qFEEIIYWatW7fOdfmBAwfytZ/Tp0+zZcsWzc/29vZMnTqV7t2752v77777juPHj+Pq6sq7776br21eJHN/GJDS5xAI7+5j7gg58th01dwRhJHIvCT6Ufr3ipKlJoca/RgJK8cZZD//+elMrsvzW6nIFB8fz507d/j+++/ZtGkTLVq0YPHixdja2ua4zeHDhxkyZAjp6eksX748y9gW+SEtFUIIIYSuDPRIaUErDXlxcHCgWrVqzJ49G0tLS3755RdWrVrF+++/n+36586dY/jw4aSmpjJz5kydKhQgj5QKIYQQRVqXLl2AnCsuly9fZvDgwcTHxzNx4kTefPNNnY8lLRVCCCGEjtTpyu9BkDkWRURERJZlmYNdRUVFMWLECAYOHKjXsaRSIYQQQuhKYU9/ZOfo0aMAlC9fXuv9+/fvM2DAAJ48ecLAgQMZPny43seS2x9CCCFEIXb+/Hm2bNlCcnJylmV//PEHn3/+OQA9e/bUvP/kyRMGDBjAgwcP6NWrl9a8IPqQlgohhBBCVwqY++Phw4dMmjSJWbNm4efnR/HixYmJieHWrVvcvXsXgIEDB9KxY0fNNtOmTeP27dvY2NiQlJTEpEmTst33hAkTsgzlnRupVAghhBC6UkCfipo1azJs2DBOnDjBrVu3OHXqFJaWlnh5efH666/Tq1cvAgMDtbaJjo4GIDk5mV9//TXHfQ8fPlwqFdkpX74MLVs1JTCwFoF1a1OtelVUKhUff7yA+fOWmjueRof2rRg9agh16vhha2vLlas3WLt2PctXrMHYQ4pY+fphXbcJqlf8sfIuB7a2qGOiSL0aTPLeLaQGn811e+uGLbBp0QGrClWwcHBCHRNFWsgtUo4eJPngLqNmz2TO8ivM2ZSaT+mfW6Xny6TEc1uY8uVKAX0qSpQowciRIwu0zffff2+ULC9NpWLYsIEMG65fr1ZjmzB+GLNnTQHgxo3bxMbFU8u/Ol98/gmtWzeje49BRvuAqWoE4PThQgDU6Wmkh4WiTkrEqmQZbBo0x6ZBcxI3f0fiL6uz2dgax1EfYR3YBIC0h6GkP36Ipas7qpp1sXB2MUmlwpzlV5izKTmf0j+3Ss8Hyj23hSWfKJiXpqNm+JMIdu7cz4wZC3mjS3+2bNlp7khaGjaoyyczJ5GWlkafvkPxrdaEuoGvUa9Be8LCHvF653Z8MHqI8QJYQNqDe8SvXET0kDeIGduf2CnvEjW4C4m//giAXbd+qOo0zLKpw3sTsQ5sQuqlc0SP6UfM6LeJ/b+hRI94i+j3upP487fGy/2M2cuvkGZTej6lf26Vnk/J57Yw5MuX9HTDvIqIl6ZSMX/eUnr2eId5c5ewb98h4mLjzR1Jy5TJo7C0tGTlqnWsX79V835QUDDjJnwMwITxwzVT2Bpa6o3LxIzrT/L+bajjYp8vSEslcf23pJzJeCTJttV/tLZT1aqHTZPWpIXeIXbOBNIfhGgtV8dEkXr+pFEyv8jc5VdYsyk9n9I/t0rPp+RzWxjy5YtabZhXEfHSVCqUzNnZidatmwKwevW6LMs3btxOVFQ0Hh7utGzR2DghEuJzrS2nnj8FgGWpMlrv23boAUDilh8gJevjTKagiPIrhNlA+fmE7pR+bpWeT+hGKhUKUKd2RuekhIQETp85n2V5amoqJ0+eA6B+/QBTx8tgYwOAOjnp+XvWNqhqBKBOTyf1zBFU1WphP2Q8jlMX4jD6Y2w7dAc7e6NHU3L5KTkbKD+f0J3Sz63S8+Wb3P7QolOlIi0tjZMnT7Jz506OHz9OSkrus+jt37+fpUuV0xNaaapUqQjA3ZD7pKWlZbvOzVt3tNY1NesGrwKQdvWC5j2r8pWxUKlQP32C7eu9cZr2ObYtO2LtF4BNg+bY9xtOsYVrsSpf2ajZlFx+Ss724jGVmk/oTunnVun58i1dbZhXEVHgG1Xnzp1jzJgx3L9/X/Ne8eLFGTNmDN26dct2mwMHDvDrr78aZAjQosjNzQWAyKdROa4TGRmlta4p2bT6D6qKPqhTkknauUnzvqVrcQAsXFyx69KHlFOHSfjxK9IfPcCqfGXsB41BVckHx7GfED1+ACQlGiWfkstPydlePKZS8wndKf3cKj2f0E2BWiru37/PoEGDCA0NxcnJCT8/P9zc3AgPD2fq1KmMHTs2z1YLkZWdnR1AtkOsZkpKylhm/2xdU7GqUBX7/hmVwcQNq0h/9LwyybMsFipr0h6GErfoo4yOmmmppN28Qtz8SagTE7D0LIlNiw5Gy6jk8lNyNlB+PqE7pZ9bpefLN3W6YV5FRIEqFd9++y2xsbF069aNw4cP88svv/D3338ze/ZsihUrxs6dO3n33XdJSEgwVt4iKTEx4y94m2f9FrJja5uxLCHROH/tZ8fSsySOE2ZjYWNL8t/7Sdq+XnuFF74MkvdthX81YaqjnpJ85A8ArGvVN1pOpZYfKDsbKD+f0J3Sz63S8+Wb3P7QUqBKxd9//02JEiWYPn265kKwtLSkW7dubNq0iapVq/LPP/8waNAgYmNj89ibyPT0WfOfay5NfK6uLlrrGpuFixuOUxZg6eZByukjxH85N8s6Lz56mnY/JMtygPTQjHuilp4ljRMUZZZfJiVne/GYSs0ndKf0c6v0fEI3BapUhIWF4e/vn23NskyZMvz0008EBgZy+vRp+vXrR2RkpKFyFmnXr98CoFxZb6ysrLJdp1LF8lrrGpOFozNOUxZgVbI0qcFnift8epZWCIC0B3ef/5DD46Tq1Ge3wyyM96CR0srvRUrO9uIxlZpP6E7p51bp+fJLnZ5ukFdRUaBvegcHh1yXOzk5sXLlSpo1a0ZwcDB9+/bl8ePHegV8GZw5e4Hk5GTs7e0JqFMzy3KVSkVgYC0Ajh8/bdwwtnY4TpyLVblKpF6/ROynU3KuMESEkx7+EABLL+9s17H0KgVA+tNw4+RFYeVXiLKB8vMJ3Sn93Co9X77J7Q8tBapUlC5dmuDg4FzXsbW1Zfny5bRt25Zr167Rp08frSdFRFYxMbEcOPAXAAMG9M6yvEeP/+DiUozw8AgOHjpivCAqaxzHzUJVtTppIbeImzsREnPvH5N87BAANs3aZl1obY1No5YApF403peCYsqvkGUD5ecTulP6uVV6vnyTjppaClSpqFOnDvfv3+fixYu5rmdtbc3nn3/OG2+8wd27dzl+/LheIV8Gc+YuJj09nUEDe9OrVxfN+/7+1Vkw/yMAFixcbrynaywscRj5f1j7BZAWFkrs7HGo42Ly3Czpt59RJ8SjeqUmtm/0AQuLjAXWNjgMGoOlmwfpsdEkH9hunNzPmL38Cmm2wpBP6E7p51bp+UTBWagLMP3b4cOHGTRoEJ06dWLhwoX52mbWrFl8//33WFhYcOnSJZ2DAjg6VNB524YN67J+wzean52cHLGzsyUuLp6EhOc9ixs36kRo6AOdjpGUqt+FP3nSSGbOmAg8n63Pr4YvVlZW7Nixn67dB5Cux7238O4+OS6zbtwKxxH/B0DagxDUUZHZrpce+YT4Lz7Wek8V0AjH0dOxsLYhPTKC9PCHWJUqi4WjE+rEBOI+m5bn/B8em64W7JfJhrHLr6hmM3Y+W5W1zrlM8bnVh3yv6M+Y+VKTQw0ZNVtxM/oYZD+O0340yH7MrUCDXzVo0IDVq1djkfnXaD5MnTqVhg0bEh0dXeBwhmRtbY2Hh3uW9x0dHXB0fN5XJKcOQ6YwZ+5igoKCGTVyMAEBNSlZ0ovzFy6zdu16li1fbdQPvsULX/xWpcpCqbLZrpf+OCzLe6mnjxAz9T3s3uiDqlptrCpUQR0dScqpwyRu/ZH0HJ4MMTRzll9hzqbkfEr/3Co9Hyj33BaWfHlSej4TK1BLhbnp01JhCvr+RWFsubVUmJshWiqEMunTUiGU/72iZCZpqZietT+ILhynZ51UrTBS8HyyQgghhMIVoSc3DEEqFUIIIYSuitCTG4YgU58LIYQQwiCkpUIIIYTQldz+0CKVCiGEEEJHRWmIbUOQ2x9CCCGEMAhpqRBCCCF0Jbc/tEilQgghhNCVVCq0SKVCCCGE0JU8UqpF+lQIIYQQwiCkpUIIIYTQldz+0FKoKhUyBr5+Sm+9Ze4IOYrZ9ZG5I+TKucPHea9kRkqeX0M+t6IoU0ulQovc/hBCCCGEQRSqlgohhBBCUaSlQotUKoQQQghdyYiaWuT2hxBCCCEMQloqhBBCCF3J7Q8tUqkQQgghdCWVCi1y+0MIIYQQBvHStVR0aN+K0aOGUKeOH7a2tly5eoO1a9ezfMUa1Grz1ziVmq98+TK0bNWUwMBaBNatTbXqVVGpVHz88QLmz1tqkgy/n73G4eDbBN8J43FUHJFxCdjZWFOppDvt6r7Cm81rYa2y0tpmxfZ/+GrnkVz3u2Xa/6hYsrgxowNybvWl1PJTejbJZ1xKz2dqL1WlYsL4YcyeNQWAGzduExsXTy3/6nzx+Se0bt2M7j0GmfUCUXK+YcMGMmz4QLMcO9N3B05y9sZ9bFRWeLo44VPak/DoOIJuPSDo1gN2HA/my5E9KOZgl2Xbkm7OlHR3zna/djbGHzhKzq1+lFx+Ss4m+UxAbn9oeWlufzRsUJdPZk4iLS2NPn2H4lutCXUDX6Neg/aEhT3i9c7t+GD0EMmXg/AnEezcuZ8ZMxbyRpf+bNmy0+QZujauyTej3+TwohHsmPkOP016m72z3+W78b0p4epE8N2HLN32d7bbdmnkx5qxvbN9lXIvZtTccm71o+TyU3I2yWci6WrDvIqIl6ZSMWXyKCwtLVm5ah3r12/VvB8UFMy4CRlDME8YPxyVyjyNN0rPN3/eUnr2eId5c5ewb98h4mLjTZ6hSyM/6vmUxdpK+xaHf0VvxvZoAcAf526YPFde5NzqR8nlp+Rskk+Yw0tRqXB2dqJ166YArF69LsvyjRu3ExUVjYeHOy1bNDZ1PMXnKwwqlnAHIDFZWfNMyLnVj5LLT8nZQPKZijpdbZBXUfFSVCrq1M7o/JOQkMDpM+ezLE9NTeXkyXMA1K8fYOp4is9XGJy79QCAauVKZLv85NUQxn3zG4M/38C4b7axeu9xwqPijJ5Lzq1+lFx+Ss4Gks9k5PaHFp0qFUFBQXz99desWbOGGzeeNzcnJyfz1Vdf8eabb9KmTRsGDRrE7t27DRZWV1WqVATgbsh90tLSsl3n5q07WuuaktLzKVVaejoPn8aw/tBZFm0+hL2tNSO6NM123VPX77H/zFVOXA1h/5lrfPHrX/xn2rdsPXLBqBnl3OpHyeWn5GwvHlPyCVMq8I2q+fPns3r1as3PCxYsYPr06XTv3p2RI0dy8OBBzbJ79+7xzz//cP78ecaPH2+QwLpwc3MBIPJpVI7rREZGaa1rSkrPpzQ//H6KBRsPar3XslYVhnVuQhVvD633PV0cGdSuAa1qV6GMhwu21iouhzzim93HOHzxFtN/2IOroz2v+lc2SlY5t/pRcvkpOduLx5R8RiZTf2gpUKXi4MGDrFq1Cnt7ezp27IhKpWLXrl3MmDEDCwsLDh48SM+ePenSpQtubm4cPXqUL774glWrVtG6dWsCAszThGVnl/GIYXJyco7rJCVlLLO3y/o4orEpPZ/SeLk6UbuyN6lp6TyIiOZJdDwnroaw68RlhnZujJXl8wa4Hs1qZdm+duXSLB3albFfb+P3c9dZsOkgzWtWwsLCwuBZ5dzqR8nlp+RsIPlMpSj1hzCEAlUqfv75ZywsLFi5cqWmgvDWW2/RrVs3PvnkE3r06MHMmTM161euXBlvb2/ef/99NmzYYLZKRWJiIgA2NjY5rmNrm7Es4dm6pqT0fErTNsCXtgG+mp/P33rAzHX7WLnnGNHxiUzt3SbPfVhYWDDqjWb8fu46IY8juRr6GN8yXgbPKudWP0ouPyVnA8knzKNAfSrOnz+Pn5+fVuWgWrVq1K5dm8TERN5+++0s27Rs2ZJy5cpx5swZ/dPq6Omz5jXXXJrQXF1dtNY1JaXnU7qaFUuxdGg3bFRWbPo7iPtPovO1XfkS7rg4ZvwFFPI40ijZ5NzqR8nlp+RsLx5T8hmZdNTUUqBKRVRUFGXLls3yvre3NwAVKlTIdruKFSvy6NGjgqczkOvXbwFQrqw3Vv8a4yBTpYrltdY1JaXnKwy8XJ3wLeNFulrN1dD8X2sqq4yPQFqacW6MyrnVj5LLT8nZXjym5DOydAO9iogCVSocHR1JSEjI8n5m85VdDve9nJ2dSU83X6mdOXuB5ORk7O3tCahTM8tylUpFYGDGvffjx0+bOp7i8xUWac+usbS0/NX6n8bGExGTMdCTl1v2Q3jrS86tfpRcfkrOBpJPmEeBKhXFixfn4cOHWd6vV68e3bt3z3G7J0+e4O7uXvB0BhITE8uBA38BMGBA7yzLe/T4Dy4uxQgPj+DgodwnnzIGpecrDEKfRHH13mMAfMp45mub7w+cQq0GJ3tb/MqXNEouObf6UXL5KTkbSD5TkcGvtBWoUuHr68vVq1eztFZkdtTMTnJyMhcuXKBcuXK6pzSAOXMXk56ezqCBvenVq4vmfX//6iyY/xEACxYuJyXFPCMyKj2fuQXffcjy7Ye5Fx6ZZdnhi7cYvnQzqenpNK1RkbKergBcvx/OrHX7uX4/XGv9pJRUvt19jDV7TwAw4LV6WWY3NSQ5t/pRcvkpOZvkMxG5/aHFQl2A6d/Wr1/PmjVrmDt3LrVqZX1ULzu//vorkyZNYvjw4QwfPlznoAAqm9J6bT950khmzpgIPJ8Nz6+GL1ZWVuzYsZ+u3QeY9TaNsfPZqnSfjbNhw7qs3/CN5mcnJ0fs7GyJi4snIeF5z+zGjToRGvqgwPsP/21KrstPXA1h8OcbAPAo5oiXq5PmkdKYhCQAapQvydJhXXFzcgDgcsgj3przPQBuTvaaicNuhj0hMTkVgK6N/ZjWp22ej5M6d/i4wL/Ti17mc5uUqv8/CEr+7Co528ueLzU51JBRsxXR9VWD7Md9yyGD7MfcClSp0MWFCxcICwujZs2alCiR/RDK+aVvpQKgU8c2jBo5mICAmlhbW3Pt+i3Wrl3PsuWrzfrBMkU+ff7hadasIbv3/JznetVeacrdu/cKvP+8KhXR8YlsPxbM8St3ufHgCU+i40hJS8fV0Q6fMl60DfChU/3qmo6XmdusP3SWszfvczssgqex8aSkpePuZE/Nit50a1KTxtUr5CufvpUKeHnPrSEqFaDsz66Ss73M+aRSYXpGr1QYkiEqFS8zff7hMba8KhXmZohKhTEp+dwaqlIhREGZpFLRxUCViq1Fo1Ih88kKIYQQOlKbv6FHUV6KWUqFEEIIYXzSUiGEEELoSloqtEilQgghhNCREm5/pKSkcOzYMQ4ePMjp06cJDQ0lPj6ekiVL0rRpU9555x1Kl867T+LNmzfp0qULycnJ1KpViw0bNhQ4i1QqhBBCiELsxIkTDBo0CIDSpUsTGBgIQFBQED/99BPbtm3j22+/pU6dOjnuIz09nalTp+o9JohUKoQQQghdKaClwsLCgo4dOzJgwAD8/f017yclJTF9+nQ2b97M2LFj2bNnD9bW2T8p9sMPP3D69Gneeustfv4570fMcyIdNYUQQggdqdMN89JHo0aNWLRokVaFAsDW1pbp06fj7OxMaGhojrOFh4SEsGjRIlq0aEHHjh31yiKVCiGEEKKIsrW11cwgntNs4f/3f/8HwPTp0/U+nlQqhBBCCB0poaUiN2lpaYSGZgwC5uHhkWX5L7/8wpEjRxgzZgylSpXS+3jSp0IIIYTQkRKe/sjN1q1biYiIoHjx4gQEBGgte/jwIfPmzaNWrVr06dPHIMeTSoUQQgihK3XukxHmV+vWrXNdfuDAgQLv8969e8ybNw+ADz74ABsbG63lH330EYmJicycORNLS8PcuJBKhQF5O7mbO0Ku7sdGmDtCjpQ+t8blKn7mjpCrV65fMHeEQquSi/5NvsZ0M6rgM8MKERMTw9ChQ4mMjKRDhw707NlTa/lvv/3GH3/8wXvvvYevr6/BjiuVCiGEEEJHhrr9ceBgwVsicpKYmMj777/PlStXaNiwIfPnz9daHhERwaxZs6hQoQJDhw412HFBKhVCCCGEztTphrn9YSjJycmMGDGCEydOUKdOHZYvX57ltsfp06d5+vQpDg4OvPPOO1rLoqOjAbh+/Tp9+/YF4Msvv8TR0TFfx5dKhRBCCFEEpKWlMW7cOP7880+qVavG119/nWtlIDQ0VPNkyL/FxcVx/PhxzX7zSyoVQgghhI6U8vSHWq1m6tSp7Nmzh8qVK7Nq1SqKFSuW7bpt2rThypUr2S47duwY/fr1k7k/hBBCCFNTG+jpD33Nnj2bLVu2ULZsWVavXo27u3keHJBKhRBCCFGI7d+/n++++w4Ab29vFi1alO16bdq0oU2bNkbNIpUKIYQQQkdKuP2R2bkSMm5f5KR06dJSqTC0Du1bMXrUEOrU8cPW1pYrV2+wdu16lq9Yg1qtNluuth1b0bJNU/xr16BESU9c3V1JSEjg2pWbbN+ym+9XrSclJdVs+TIptfzMnc8uoAaOrRphX9cP64plsbS3Je1pNIlnLxH10zYSjp/Lso1Ntco4tW6Cfb2a2FQpj6WTI2nRsSQFXyN6w07iDvxjtLzZkXNrGKMnv8/QMRnTUC+as4IVn600cyLll53S8+VGCU9/dOvWjW7duhlkXw0aNMixv0V+WKiVfsZeoLIprdf2E8YPY/asKQDcuHGb2Lh4/Gr4YmVlxbbf9tC9xyC9LmB9Br/atHMt9RoGkJiYxKOwxzyNiMSrpCelvEsAEHTmIv/tOpjo6Bidj6Hv4FfGLj99GTNfboNf2TesTelVGaPWqdPSSLl7n/SERGzKl8bS0QGAiBU/ErHkO802qrKlqLBnjebnlJAHpEXFYF22FFYuzgBEb9nLow8/g3xk1nfwq5f53Bpy8KvKVSuw9Y+fsLHNeITPEJUKfQe/epnPbWpy9k82GFJIvdxHwsyvsicMN06FOb00E4o1bFCXT2ZOIi0tjT59h+JbrQl1A1+jXoP2hIU94vXO7fhg9BCz5Vv3/Wbe7DyA6uUa0jSgA53b9KaBXxu6tO3D/dAw/OvUYPyHI82WT+nlZ9Z8FhYk3wnl0cdLuNXkTe52eod7PYZzs3FPIr7+GQD39/vg8GqDFzaxIPXRE8IXfMut5r250+5/3HtzBLca9+TxJ8tQp6dTrGtbXHp3Nk7mF8i5NZwZC6eQkprKkT+PmzsKoPyyU3q+/FCrDfMqKl6aSsWUyaOwtLRk5ap1rF+/VfN+UFAw4yZkDBE9YfxwVCrz3BHauG4rRw+fJDVV+xbHmZNBzPzwUwDadWxpjmiA8svPnPkSg65w9z/vEL1+O+nRsc8XpKQS8flq4p79A1OsZwfNotSwx9xpP4DIVb+QFv5CC5JaTdRP24jesDNjmx7PtzEWObeG0fPtLtRrFMCyBd/y4P5Ds2bJpPSyU3q+/FCnWxjkVVS8FJUKZ2cnWrduCsDq1euyLN+4cTtRUdF4eLjTskVjU8fL0/WrtwCwd7A3y/GVXn7mzqeOi4e0nHtrxf9zGgCb8s9v36mTU1AnJuW8zeFTAFhX0O+WX17MXXZ5UXq+TG7FXRn3fyO4dvkGa7780Ww5XqT0slN6vvySSoU2g1cq9u7dq3m0RSnq1M7o/JOQkMDpM+ezLE9NTeXkyYyOdPXrB2RZbm5169cC4MK5YLMcX+nlp/R8ls/ur6cn5VyJ+DeLZ9uok5KNkimT0stO6fkyTZk5Bjd3V6ZPnEdqav5HHzQmpZed0vMJ3Ri8UvH9998zZ84cQ+9WL1WqVATgbsj9HIcbvXnrjta65mZpaUlJ7xL0HdiLqTPGEhcbz9yZX5gli9LLT+n5HNs1AyDxTP4rhU7tm2dsc/qiUTJlUnrZKT0fQKNm9ejSsyNbN+zkxLNWKSVQetkpPV9+SZ8Kbfm+UXXr1q18rZeYmAjA7du3tXrsVqxovovCzc0FgMinUTmuExkZpbWuuQx6720+mj1R673d2w+wYM5Srl66bpZMSi8/Jecr1rMDdtWrok5OJvK7Lfnaxr5xAE5tmgDwdNUvxoyn6LJ78ZhKzWdja8PHCyYTHRXD3Omfm/z4uVF62Sk9X34VpVsXhpDvSkWHDh2wsMh/4XXo8LyDmYWFBcHB5mm6B7CzswMyZm/LSdKzZmb7Z+uaS9iDR5w4ehqVtYrSZbzxKuGR8ZfQlQ4snLOM9HTTj7Si9PJTaj7balXwmPw+AE8WryU1JO9HA1WlPCk5P6NSGfnTNhJP6feoaF6UWnaZlJ5v6JhBVKhUjo8nzuPJY/0e2TY0pZed0vMJ3RSoS62FhQUlS5bMdZ3w8HBSUlIoVcpwz37rK7P15N/Tv77I9tk97IRn65rLjq172bF1r+bn2nVrMuezaYwYOwRXNxemjvvE5JmUXn5KzKcqXYJSK2ZgaWdLzPbfiVy1Mc9tLF2cKfXVLKzcXYk/do7weV8bPacSy+5FSs5XuWoF3hnWlwvnLvHT6rzPr6kpuexA+fnySylzfyhFvisV9evX5/jx41SpUoXp06fj7e2d7Xp9+/bl5MmT/P777wYLqa+nz5rXXHNpQnN1ddFaVynOnjrP/3oN5a/Tu/hv/x4s/3wloff0GwynoJRefkrLZ+XhhvfKuai8ihN38BgPpyzIcxsLBzu8v5yJbZXyJF64yoNhH0FKitGzKq3s/k3J+T6aPwkrlRXTx89V5KiPSi67F4+p1Hz5pYRhupUk35WK7777jp9//pkFCxbwn//8hw8++IC33367QLdEzOX69Yz+IOXKemNlZZVtp6BKFctrraskD8MeE3zhMgGBtaju52vySoXSy09J+SxdnPH+dg425bxJOH6OsA8+gbyeBrC2ptTS6djVqkbS9TvcHzIVdXyCUXNmUlLZZUfJ+arX9EWtVrPih4VZljk7OwEweEQ/3h7Uk7DQh3Rv29+k+ZRcdi8eU6n5hG4K9PTHW2+9xfbt2wkMDGTWrFn06tWLq1evGiubwZw5e4Hk5GTs7e0JqFMzy3KVSkVgYMZjm8ePK6f39otUVhn1PyuVlcmPrfTyU0o+TWuDT0USg65wf+hHeT8SamVJyUVTcWhYh5S797k/aBLpkdG5b2NASim7nCg9n0qlwtPLI8vLzj6jD4CTkyOeXh64FXczeTall53S8+VXutrCIK+iosCPlJYsWZKvv/6auXPncufOHbp168aiRYty7WxjbjExsRw48BcAAwb0zrK8R4//4OJSjPDwCA4eOmLqeHkqU9aban4+AFy6oPtEL7pSevkpIt+LrQ3XbnP/3fy1NpSYPQ6nVo1IfRhO6KBJpJm4s58iyi4XSs4XWKUlPp6B2b42//wbkDH3h49nIK3qvm7SbKDssgPl58svtdrCIK+iQudxKt544w127txJq1at+Oqrr+jcuXOuU66a25y5i0lPT2fQwN706tVF876/f3UWzP8IgAULl5NigvvY/1azVnXGTBpKufJlsix7tXUT1v6yAmtraw7s/ZM7t++ZPB8ou/zMns/SkpILJ+PQsA7Jd+9zf9Bk0qPynvjNY8r7OHduTVpEJKGDJpEaap6hneXcFl1KLzul5xMFZ5BZSvfu3cuMGTN48uQJtra2JCUlcenSJUPk06LvLKWTJ41k5oyMx/X+PRvejh376dp9gF6PbOo6S2nDJoFs+G01AI/CHvPg/kOsbawpXaYULq7FgIwOm/17DeVpRKTO+fSdpdTY5acvY+bLbZZSp44tKLlgMgDJt++R9iQy2/XSwiMI+2AWAHa1qlFm3ecApDx4ROr9RznuP7Tv2Dzz6TtL6ct8bg05S2mmuUs+ottbnRUxS+nLfG5NMUvpZZ+OBtnPK1d3GmQ/5maQWVratm1LgwYNmD17Nr///nuujwiZ05y5iwkKCmbUyMEEBNSkZEkvzl+4zNq161m2fLXZPliXLlzlo0lzaNK8IT6vVKZy1YpY21gTGRHJqRPn2P7rHrZs2J7jqHOmotTyM3c+CxtrzX/bVCgDFbK2OAGkhIZlu411KS+sS3kZJVt+ybktupRedkrPlxcFPvhjVgZpqTAVfVsqjE3XlgpT0bel4mWWW0uFEujbUvEyM0ZLhSHp21LxMjNFS0Vw5U4G2U/1GzsMsh9zeylmKRVCCCGE8Sl3knohhBBC4YrS46CGIJUKIYQQQkdF6XFQQ5DbH0IIIYQwCGmpEEIIIXRUeB51MA2pVAghhBA6kj4V2uT2hxBCCCEMQloqhBBCCB1JR01tUqkQQgghdCR9KrTJ7Q8hhBBCGIS0VAghhBA6ko6a2qRSYUAyt0bRVev2FXNHyFXM90PMHSFHZQf/aO4IQhiN9KnQJpUKIYQQQkfSUqFN+lQIIYQQwiCkpUIIIYTQkTz8oU0qFUIIIYSO5PaHNrn9IYQQQgiDkJYKIYQQQkfy9Ic2qVQIIYQQOko3dwCFeekqFR3at2L0qCHUqeOHra0tV67eYO3a9SxfsQa1AsZblXxFL1/58mVo2aopgYG1CKxbm2rVq6JSqfj44wXMn7fUJBl+vxTC4WsPCA59wuOYBCITkrGztqKSpwvt/MrxZr2qWKustLY5cfMhv18K4WLoEx5ExRMZn4S1lSXlizvTsloZ+jR6BUdba6Nnb9CwLh06taZh40CqVq2EvYMdEU+ecuL4WVZ+/QN//3XM6BkKYvTk9xk6ZhAAi+asYMVnK82cSJmfi8KUT+SfhboQnTGVTWm9tp8wfhizZ00B4MaN28TGxeNXwxcrKyu2/baH7j0GmfUClnzKzWer0v0fz/nzpzFs+MAs7xuyUhG+ekCuy//37T7O3n2MjcoST2d7XOxtCY9N4FF0AgDVvd35sn8ritnbaLaZsvEfdgbdRmVpgaezPW6OdjyNTyIsKg61Gkq7OfLtgDaUcnXM9dj6DH7V7NVGbPltLQBpaWncunmH+LgEKlUuj5OzEwAL5i9j7idf6HwMd9tiOm/7b5WrVmDrHz9hY5tRjoaoVNyMeqDX9i/z5zY1OdSQUbP1Z8meBtlP87BfDLIfc3tpOmo2bFCXT2ZOIi0tjT59h+JbrQl1A1+jXoP2hIU94vXO7fhgtPlGJZR8RTdf+JMIdu7cz4wZC3mjS3+2bNlp8gxd61bmmwGtOTz1TXZ80IWf3mvP3nFd+W5wW0oUcyD4fgRLD5zT2qZVtTIs69uCv6f2ZNfYN/jpvfbsGtOFjcM64VPCldCnccz67YRRc1tYWHDzxm3Gf/ARPhUb0LBue1o170rVCg1YtOBLAMZNGEbb9i2MmiO/ZiycQkpqKkf+PG7uKICyPxeFIV9+pKsN8yoqXppKxZTJo7C0tGTlqnWsX79V835QUDDjJnwMwITxw1GpzHNHSPIV3Xzz5y2lZ493mDd3Cfv2HSIuNt7kGbrUqUS9iiWwttL+yPuX9WBs+wAA/rh0T2tZmxrlaFLVGztr7TKr7OXCtDcaAHDkxgOSUtKMlvv0qXM0CuzA6pXriIqM1ryfkpLCrBmfsW/vIQD69n/TaBnyq+fbXajXKIBlC77lwf2H5o4DKPtzURjyiYJ7KSoVzs5OtG7dFIDVq9dlWb5x43aioqLx8HCnZYvGpo4n+fSk9HxKV9Ezo/k/MSU1/9t4ZGyTlq4mOc14lYrYmDjSctn/oT8OA1C5SkWjZcgPt+KujPu/EVy7fIM1XypjrhOlfy6Uni+/0rEwyKuoeCkqFXVqZ3T+SUhI4PSZ81mWp6amcvJkRtNv/foBpo4n+fSk9HxKdy4kHIBqpdwLvE0ZNyec7WzyWNt4bG1tAUhMSDRbBoApM8fg5u7K9InzSE01XiWrIJT+uVB6vvxSY2GQV1Fh0Dal3bt38+eff/Lo0SMcHBzw8/Oje/fuFC9e3JCHKbAqz/6KuRtyP8e/em7eukNrmmnWNSXJpx+l51OitPR0wmMSOXjlHov3ncXeRsWI12rluo1areZJbCLHbobx+d6zqCwtNLdOzKXLG+0BOHbstNkyNGpWjy49O7J1w05O/GO+HP+m9M+F0vPllzxSqq1AlYrp06dTu3Zt3njjDa33o6KieP/99zlz5oxWL919+/bx1Vdf8fnnn9OsWTODBNaFm5sLAJFPo3JcJzIySmtdU5J8+lF6PiX54Z/LLNit/Q9fy2plGNbKnyolXLPd5vdLIYxZ95fWe3UreDH/zabUKe9prKh56ve/XvjXrkFSUjJfLV9jlgw2tjZ8vGAy0VExzJ3+uVky5ETpnwul5xO6KdDtj59//pljx7I+Ez5hwgROnz5NsWLFGDBgADNmzGDs2LH4+fkRFxfH6NGjuX//vsFCF5SdnR0AycnJOa6TlJSxzP7ZuqYk+fSj9HxK4lXMgdrlPPErXZziThllceLWQ3adv0NaevZ/c7na21K7nCf+ZYrjVcweCwu4GPqE7eduFagfhiH516rOrHlTAZgzcxG3b4WYJcfQMYOoUKkci2Yv58njCLNkyInSPxdKz5dfcvtDm963P4KCgjh06BAlSpRgw4YNlChRQrNs8ODBTJ06lU2bNvH9998zceJEfQ+nk8TEjPutNjY53/u1ffZceUKi6e/NSj79KD2fkrT1K0dbv3Kan8+HhDPzt+Os/PMi0QlJTO1cP8s2ARW8WPPOa5qfbz6OYs72k2w6eZ2wyDiW9WtpkuyZypUvw08bvsLe3o6NG7axdLF5BpeqXLUC7wzry4Vzl/hp9UazZMiN0j8XSs+XX3L7Q5veHTX/+ecfLCws+OCDD7QqFJkmT56Mi4sLf//9t76H0tnTZ81rrrk0obm6umita0qSTz9Kz6dkNct6sPTtFtioLNl08gb3I+Py3KaSpwuL+7xKcSc7Dl9/wJk7j0yQNIOXlwcbt66mZKkS7N39B8Pfm2SyY//bR/MnYaWyYvr4uYoc9VHpnwul5xO60bul4sGDjNHeGjRokO1yJycnqlWrRlBQkL6H0tn167cAKFfWGysrq2w7BVWqWF5rXVOSfPpRej6l8yrmgG9JN87fe8LVsKd45zFCJoC9jYrACl7suXCXSw+eUqe8l9Fzurq5sHHraipVKs/hv44xsN9IUlPNc/sFoHpNX9RqNSt+WJhlmfOz0T4Hj+jH24N6Ehb6kO5t+5s0n9I/F0rPl1/SUqFN75YKa+uM4Yvd3XN+HM3Dw8OsH/4zZy+QnJyMvb09AXVqZlmuUqkIDMzo+X78uOl7b0s+/Sg9X2GQ9mxIv7QCDO2n2SbN+F+rjo4O/LzxG6rX8OX0qSD69HqPxMQkox83LyqVCk8vjywvO/uMPgBOTo54enngVtzN5NmU/rlQer78kj4V2gpcqQgPD+fEiROaV6bMFovsPH36FBcX8/XejYmJ5cCBjN7rAwb0zrK8R4//4OJSjPDwCA4eOmLqeJJPT0rPp3ShT2O5GvYUAJ+SrvnaJiYxmRO3MkaN9C1l3H8wbWys+X7dCgLr1eZS8FXe7DaI2Ni8b9MYW2CVlvh4Bmb72vzzb0DG3B8+noG0qvu6yfMp/XOh9HxCNwWuVPz999/069dP8/rxx4zR444ePZrt+ikpKVy4cIHSpfWbDExfc+YuJj09nUEDe9OrVxfN+/7+1Vkw/yMAFixcTkpKiuSTfEVK8P0Ilv8exL2I2CzLDl+7z/DvD5KarqZpVW/KujsD8Cg6nvk7T3H9UWSWbYJCwhn23R9EJSRTtYQrdSsY79aHpaUl36z+nOYtGnHz5h16dBmQ6yOIQpvSPxdKz5cf6RaGeRUVBZqldPLkyTkuq1atGv369cvy/vbt2xk3bhz9+vVjypQpuqV8Rt9ZSidPGsnMGRlPoPx7NrwdO/bTtfsA0nN4rM4UJJ9y8+kzS2nDhnVZv+Ebzc9OTo7Y2dkSFxdPwgsjQTZu1InQUN1mpMxtltITtx4yePUBADyc7PAq5kBqWjoPouKIScz4sq5R2p2lb7fAzTGj2T70aSydFm0DwMXeBm9XR9TAw6h4nsZn3HYo6+7E8n4tNRWRnOgzS2m3Hp34etUiAG5cv8XjHB7bfBj2iEH9R+l0DEPOUppp7pKP6PZWZ0XMUvoyf25NMUvp1pL/Nch+uoT9ZJD9mFuBOmrOmTOnwAeoXLkyS5cuxdfXt8DbGtqcuYsJCgpm1MjBBATUpGRJL85fuMzatetZtny1WT9Ykq/o5rO2tsbDI2ufI0dHBxwdHTQ/W1lZGeX4viXdmNCxLsdvhnHjURS3w6NJSUvH1cEW/7IetPUrTyf/CqhemGzMw8meD1+vz/GbYVx58JSQp7EkJqfibG9D/YolaFmtDF3rVs4y2Zihvfi4YeUqFXOc4+PunXvZvi+U+7koLPlEwRSopcLc9G2pEEJX+rRUmEJuLRXmpk9LhSkYo6XCkPRtqXiZmaKl4lcDtVS88TK2VAghhBDiOWlH0SaVCiGEEEJH6RZFqJelAbwUU58LIYQQwvikpUIIIYTQkVI6JV68eJF//vmH8+fPc+HCBUJDM/qTHDhwgDJlyuS67fnz51m7di0nTpzgyZMnODs7U758edq0acM777xToBxSqRBCCCF0pJQ+FcuWLePAgQMF3m7VqlV8+umnAPj7+1O3bl0iIiK4evUq69evl0qFEEII8bKpXbs2Pj4++Pn5UbNmTbp160Z4eHiu2+zatYt58+ZRpkwZli9frjX0Q1paGhcvXixwDqlUCCGEEDpSymiYQ4YMKdD6ycnJfPLJJ1hZWbFs2bIsY0lZWVnh7+9f4BxSqRBCCCF0lF5IJwPbt28f4eHhNG/enFdeecVg+5VKhRBCCPGSOXIkY5K2Jk2aEB0dzfbt27l69SoqlYrq1avTvn17HBwc8thLVlKpEEIIIXSklKc/Cur69esAxMfH07FjRx4/fqy1/LPPPmPp0qXUrl27QPuVSoUQQgihI0P1qWjdunWuy3V5siM3kZGRQMZTIyVKlOCbb74hICCAsLAwFi9ezJ49e3j33XfZsWMHHh4e+d5voapUKH3+BaVLSlXu9MFCP859vzZ3hBzFHlpg7gi5cnp1nLkjCGFymRO1paens2LFCk1HzSpVqvDFF1/QrVs3goOD+fHHHxk1Kv8zABeqSoUQQgihJIYap8LQLRF5cXR0BKBWrVpZnvywsLCgR48ezJgxg+PHjxdovzJMtxBCCKEjtYFepla6dGmt//+3zFE48xrr4t+kUiGEEELoKN3CMC9Tq169OgBRUVHZLs/sc1HQJ0CkUiGEEEK8ZDI7hp4/f564uLgsy48ePQpAjRo1CrRfqVQIIYQQOko30MvUfH19adGiBZGRkXzyySekpDzvyP/HH3+wdetWLC0teeuttwq0X+moKYQQQuhIKROKHTx4kOXLl2t+zrytMXz4cGxsbAB49dVXGTZsmGadTz75hN69e7N582aOHTtGjRo1CAsL4/z586jVaiZOnIifn1+BckilQgghhCjkIiIiOHfuXJb3L126pPnvSpUqaS3z9PRk8+bNrFixgv379/PHH3/g4OBA06ZNGThwII0bNy5wDgu1Wl1oBgRzdKig87bly5ehZaumBAbWIrBubapVr4pKpeLjjxcwf95Sw4VUcD5DjFPRoX0rRo8aQp06ftja2nLl6g3Wrl3P8hVrUMKlZKx8+oyR8rKf2/yMU/H7qUv8ff46wbfu8zgyhsjYeOxtrKlU2pN29f14s1Ug1qqc/wbac/wCv/55hst3w4iJT8TVyYGqZbxoW78GXZsH5HpsQ4xT8bJ+LpSeLzU51IAps/dl2bcNsp/3Qn4wyH7M7aVpqRg2bCDDhg80d4wcKT0fwITxw5g9awoAN27cJjYunlr+1fni809o3boZ3XsMMusXlFLzybnN29rd/3D2Wgg2Kis83ZzxLVeSx5ExnLt+j3PX77H9n3N8Nb4fxRzttbZLTkll/PJfOHjmCgBlPN0oVdyFJ1GxHL14k8jY+DwrFfoyd9lJPvNSyu0PpXhpOmqGP4lg5879zJixkDe69GfLlp3mjqRF6fkaNqjLJzMnkZaWRp++Q/Gt1oS6ga9Rr0F7wsIe8XrndnwwumBT774s+eTc5q1b8wC+ndiff76cws5PR/PTR0PYt2gs3384iBLuxQi+/YClm37Pst20lb9y8MwV6vqWZ+uc4ez4dBQ/fTSEPZ+N4cAX4xjRI/ehj/WlhLKTfEJJXppKxfx5S+nZ4x3mzV3Cvn2HiIuNN3ckLUrPN2XyKCwtLVm5ah3r12/VvB8UFMy4CR8DMGH8cFS5NFG/rPnk3OatS7M61KtWEWuVldb7/lXKMu6tdgD8fvqy1rLDQdfYdfQCFUt5sHzs21QopT0/gXsxRxr7VTFaZlBG2Uk+8yqsT38Yy0tTqRC6c3Z2onXrpgCsXr0uy/KNG7cTFRWNh4c7LVsUvGOPvpSeT8kKQ9lVfFZZSEzW7jfyw96M5+gHd26OnY3p5wVSetlJPtMorCNqGovBKxX37t3j7NmzXL16ldTUVEPvXphBndoZnacSEhI4feZ8luWpqamcPJnR67h+fePev86O0vMpWWEou3M3QgCoVr6U5r3E5BSOX7qFhQU0q1WVE5du8dHKrQyet5YxS9bzw54jxCUkGTWX0stO8glzKFCb0qNHj7h37x4VK1bEzc1Na9mmTZtYtmwZDx480LxnZ2fHm2++yZgxY7C1tTVMYmFyVapUBOBuyH3S0tKyXefmrTu0pplmXVNSej4lU2rZpaWnEx4Zw8EzV/jil/3Y21ozsufz/hFX74aRmpaOl5szq3ceZtWOv7W2P3DqEmt2/cPSD/7LKy9URgxJqWWXSfKZhjmG2FayAlUqFi9ezKZNm9i7d69WpeKLL77gyy+/1PTQLV68ONHR0SQkJPDdd98RFBTE2rVrNQNwiMLFzc0FgMin2Y8RDxAZGaW1rikpPZ+SKa3sfthzhE/X7dF6r2XAKwzr1pKqZUpo3nscFQtARHQcq3b8zau1fRjTqy2lPV25cvchn6zdzqU7Dxi9+Gc2zxqKg53h/6hRWtn9m+QzjaLUH8IQCnT749SpU1SpUoWyZctq3rt16xZfffUVNjY2TJ48mTNnznD48GGCgoL4/vvv8fX15ezZs3z33XcGDy9Mw87ODoDk5OQc10lKylhm/2xdU1J6PiVTWtl5uRWjdtWy+FUqTfFiGVMzn7x0i91HL5CW/vzrO+FZptS0dMp4urFweC8qlPLAWqXCr1Jpln7wX+xsrHnwJIpf/zprlKxKK7t/k3ymIR01tRWoUvHw4UMqVKig9d6+fftIT09n0qRJ9O/fH3v7jOfILSwsqFevHl9//TVOTk5s377dYKGFaSUmJgLk2tJka5uxLOHZuqak9HxKprSya1u/BmunDuLHaYP5ffF4vv+/d/D2cOXb7X8x5/vnj+LaWj9vZH2zVb0sT414uDrTvkHG8ML/nL9ulKxKK7t/k3zCHApUqVCr1Vhaam8SGhqKhYUF7dq1y3YbLy8v6tSpw+3bt3UOKczr6bPmSddcmiBdXV201jUlpedTMqWXnX/lMiwd0wcblRWbDp7ifngkgNYgWBX/9ShppkreGe+HPtvG0JRedpLPNOTpD20FqlSULVuW8+e1e+lmzrX+4gxn/5aamqro54xF7q5fvwVAubLeWFlZZbtOpYrltdY1JaXnU7LCUHZebsXwLVeSdLWaqyFhAFQoWVyz3MY6+9yZw3qnpxuncVnpZSf5TCPdwjCvoqJAlYrWrVvz4MEDfvzxR817jRo1Qq1Ws3Xr1my3uXv3LmfOnMHHx0e/pMJszpy9QHJyMvb29gTUqZlluUqlIjCwFgDHj582dTzF51OywlJ2mf0pUtMy/r+Euwsl3YsBcO/x02y3CX32vpdbMaNkUnrZST5hDgWqVAwaNAgPDw9mz57Np59+ypMnT2jevDktW7Zk8eLFLFiwgDt37pCamkpERAQ7d+5k4MCBJCYm0qdPH2P9DsLIYmJiOXDgLwAGDOidZXmPHv/BxaUY4eERHDx0xNTxFJ9PyQpD2YU+fsrVkIcA+JYrqXn/tXo1APjtcNaZGZOSU9hz/AIA9asZ53FEpZed5DMN6aiprUCVCicnJ1avXk3x4sVZtWoVzZs3p2vXrpp+FitXrqR9+/bUrFmTJk2aMHbsWO7du0fv3r3p1KmTUX4BYRpz5i4mPT2dQQN706tXF837/v7VWTD/IwAWLFye622wlzmfkpm77IJv32f5lj+49ygiy7LDQdcY9tmPpKal08y/KmW93DXL/tehMQ52Npy9FsI32/7U3OZITE7hk7XbeRwZSzFHO7q3qGuU3GD+spN85id9KrTpNPV5TEwMS5Ys4ddffyU6OjrH9apXr867776bYyfOgtJn6vOGDeuyfsM3mp+dnByxs7MlLi6ehITnPYsbN+pEaOiD7HZhVKbIp+/02JMnjWTmjInA89kE/Wr4YmVlxY4d++nafYDR7l+bO58+U5+/7Oc2r6nPT1y6xTvz1gLg4eJECfdipKSm8eBJFDHxGeVTo6I3y8b0wc3ZUWvbg2euMG7ZBlJS0yhezJGSxV24G/aEmIQk7GysWTSyV57zf+g79fnL/LlQej5TTH0+p7xhpj6ffKdoTH2uU6UiU3JyMhcuXOD69etERkaSnp6Oo6MjpUuXpnr16pQsWTLvnRSAPpWKZs0asnvPz3muV+2Vpty9e0/n4+jKFPn0/YcHoFPHNowaOZiAgJpYW1tz7fot1q5dz7Llq836xWTsfPpUKl72c5tXpSI6LoHf/jnH8eBb3Ah9xJPoOFJS03B1sse3XEna1qtBp8b+qHLozHct5CHfbv+Lk5dvExkbj5uzIw1rVGJQp6ZU9PbMM5++lQp4eT8XSs9nikrFrPKGubU/9c6Pea9UCOhVqTA1fSoVwjD/8Lys9KlUmIKSz21elQpzM0SlQiiTKSoVMw1Uqfi/IlKpkFlKhRBCCGEQMniEEEIIoaNC09RvIlKpEEIIIXRk/h4pyiKVCiGEEEJHRWk0TEOQPhVCCCGEMAhpqRBCCCF0lC69KrRIpUIIIYTQkVQptMntDyGEEEIYhLRUCCGEEDqSpz+0SaVCCCGE0JH0qdAmtz+EEEIIYRCFqqVCyfMbCCGyp/S5NRLu/2XuCLmy925m7ggiF9JOoa1QVSqEEEIIJZE+Fdrk9ocQQgghDEJaKoQQQggdSUdNbVKpEEIIIXQkVQptUqkQQgghdCR9KrRJnwohhBBCGIS0VAghhBA6UssNEC0vXUtFh/at2LPrZx6FXSDq6TWOH9vNsKEDsLCwMHc0QPLpS4n5ypcvw/8GvMXSZXM4enQXUdHXiYu/zYSJw82WKTtKLLsXKSnf4q/X4tekA35NOvDVmnU5rnf2wiVGTPyYZp16UbdlF17vM4QvV/9EUlKyCdMqq+yyo/R8uUk30KuosFCr1YWmmqWyKa3X9hPGD2P2rCkA3Lhxm9i4ePxq+GJlZcW23/bQvccgzFkckk+5+WxV1jrnmj9/GsOGD8zy/scfL2D+vKU67/dF+g4M9zKf24IOfnXj9l16/G8YKSmpAIwY3I93/9c7y3rb9/zO1FkLSUtLp4RncdzdXLl28w6pqan4VfNh9dJ52NvZ5Xk8fQe/epnPbWpyqCGjZmt4hV4G2c/S2+sNsh9ze2laKho2qMsnMyeRlpZGn75D8a3WhLqBr1GvQXvCwh7xeud2fDB6iOSTfAYX/iSCnTv3M2PGQt7o0p8tW3aaJUdOlFx2SsunVquZMX8JKpWKBnVr5bhe6IOHTJv7OWlp6YwZOoj9W77nl9VL2bn+WyqWK8OFS1f5bNkqo+dVUtkVxnz5kY7aIK+i4qWpVEyZPApLS0tWrlrH+vVbNe8HBQUzbsLHAEwYPxyVyjzdTCRf0c03f95SevZ4h3lzl7Bv3yHiYuNNniE3Si47peXb9NseTp27wPsD/ktJL88c11v900aSk1NoXD+AgX16aJrxvUuWYMaUDwD4ZdsuwiOeGjWvksquMObLD7WBXkXFS1GpcHZ2onXrpgCsXp31/ufGjduJiorGw8Odli0amzqe5NOT0vMpmdLLTkn5Ip5GsmjFKipXKEffXl1zXE+tVnPg0D8AdPtPuyzL69SsTsXyZUlNTeWPv44YLa+Syi47Ss8ndPNSVCrq1PbD1taWhIQETp85n2V5amoqJ0+eA6B+/QBTx5N8elJ6PiVTetkpKd+8xV8TFR3Dh+OGYZ3LX84PHj7i8ZMIAOr4V892nTo1M94PCr5i+KCZx1BQ2WVH6fnyS25/aCtQpeJ///sfy5cvJzTU+J1fDKlKlYoA3A25T1paWrbr3Lx1R2tdU5J8+lF6PiVTetkpJd/Rk2fYsfcP/tOuFfXq+Oe67p2Q+wDY2Fjj5VE823XKeJcEMn4vY1FK2eVE6fnyS57+0FagSsXRo0dZsmQJr732Gv369WPLli3Exyvr/nB23NxcAIh8GpXjOpGRUVrrmpLk04/S8ymZ0stOCfmSkpKZ8elSnJ0cGTf8nTzXj46JBcDZySnHRyKLFXPSWtcYlFB2uVF6PqGbAt/+cHR0BOD48eNMmTKFJk2aMGnSJI4cMd69QX3ZPXtsKzk552fDM58bz88jXoYm+fSj9HxKpvSyU0K+r9au4+69+4wc0h8Pd7c81096ltXaOudbJDbWGY8oJyYlGSZkNpRQdrlRer78Uhvof0VFgSsVr732Gr///jujR4+mYsWKJCQk8OuvvzJw4EBatmzJ559/zu3bt40QVXeJiYkA2NjY5LiOrW3GsoRn65qS5NOP0vMpmdLLztz5bty+y+qfNlLdtwq9unbK1za2z7JmjmORneSUjHFF7Gxt9Q+ZA3OXXV6Uni+/5PaHNp06apYsWZL33nuPnTt3smHDBnr37k2xYsV48OABX331FR06dKB3795s2LCBmJgYQ2cusKfPmtdcc2lCc3V10VrXlCSffpSeT8mUXnbmzvfJgmWkpaXzf+OGY2mZv6/LYs4ZtzZiYmNzHLQpOjpWa11jMHfZ5UXp+fJLWiq06f3wr7+/P/7+/kyZMoU//viDLVu28Ndff3HmzBnOnj3LrFmzaN26NW+88QbNmzc3ROYCu379FgDlynpjZWWVbaegShXLa61rSpJPP0rPp2RKLztz57t87QYWFjB84sdZlsXGxgGw8odf+GnTb5T08mD9ysWUL+sNQHJyCo/Cn1DC0yPLtvfuhwEZv5exmLvs8qL0fEI3Bnuk1NramrZt27JixQr++usvpkyZQrVq1UhKSmLnzp289957hjpUgZ05e4Hk5GTs7e0JqFMzy3KVSkVgYMboeMePnzZ1PMmnJ6XnUzKll50S8qWlpfMk4mmWV2bfifiEBJ5EPOXps06FpUp44VE8o+/FmaDgbPd55nzG+/7VfY2SGZRRdrlRer78ktsf2owyToWbmxv9+vVj8+bNbN++nYEDB+LhkbW2bioxMbEcOJAxvv+AAVnH6O/R4z+4uBQjPDyCg4dM3+FU8ulH6fmUTOllZ+58R/Zs5MLhXdm+unRoA2TM/XHh8C72bloLgIWFBa2bZwzWtHn7niz7PHM+mFt3QlCpVLRs2tDgmTOZu+zyovR8+ZWuVhvkVVQYffCrKlWqMGHCBA4ePGjsQ+VqztzFpKenM2hgb3r16qJ539+/OgvmfwTAgoXLSUnRb2ImySf5Chull53S82VnwH97YG2t4p/jp1n140ZN34r7YQ+ZNnsRAN07t8OjuLtRcyi97JSeTxRcgWYpfeWVV+jatStz5swxZqYc6TtL6eRJI5k5YyKQdTa8HTv207X7ANLTzdcQJfmUm0+fWUobNqzL+g3faH52cnLEzs6WuLh4EhKe92pv3KgToaEPdDqGvrOUvszntqCzlGaa+slCtu7an+MspVt37ef/Zi8iPT3rLKXVfauyZtl8HOyNP0vpy3xuTTFL6dvluxlkPz/c2WyQ/ZhbgSoVx48fx8PDg0qVKhkzU470rVQAdOrYhlEjBxMQUBNra2uuXb/F2rXrWbZ8tVk/WJJP2fn0qVQ0a9aQ3Xt+znO9aq805e7dezodQ99KBby859ZYlQrIuNXx7ffrOXv+EgmJiZQuWYIOr7VgUJ+emscl86JvpQJe3nNrikrFf8vnPA9MQfx0Z4tB9mNuBapUmJshKhVC6EKfSoUpGKJS8bLStVJhKoaoVLyspFJhesqdT1YIIYRQuKI0xoQhSKVCCCGE0JH5bx4py0sx9bkQQgghjE9aKoQQQggdpSvo9sfNmzf55ptvOHbsGI8ePUKlUlGuXDnatGnDwIEDcXIy3rDwmaSlQgghhNCRUub+OHnyJF27dmXz5s1YW1vTqlUr6tWrR2hoKMuWLaNHjx5ERkbq/wvnQVoqhBBCCB0ppU/F9OnTSUxMZOjQoYwcORILCwsAIiMjGThwIBcvXuTbb79l3LhxRs0hLRVCCCFEIfb06VOuXbuGtbU177//vqZCAeDq6srAgQMBCAoKMnoWaakQQgghdKSEoZ6srfM3jo6bm5uRk0hLhRBCCKGzdNQGeenDycmJ2rVrk5KSwooVK7QqOpGRkaxatQqA7t2763Wc/JCWCiGEEKKQ++STT3jnnXdYvnw5O3fuxNfXl8TERE6dOoW9vT3z5s2jefPmRs8hlQohhBBCR4bqqNm6detclx84cCDX5VWrVmX9+vWMHDmSc+fOcfv2bc2yRo0aUaVKFUPEzJNUKgzoPe+m5o6Qq23RweaOkKP41CRzR8hVZGKcuSPkytvJuFNoF2WlKrU3d4RcvepVw9wRchSSFGHuCGanlGG6jx49ysiRI/Hw8ODbb7+lVq1aJCQkcPDgQT799FP+/PNPli1bRrNmxp1LRioVQgghhJnl1RKRm6ioKEaNGkVycjLffvst3t7eABQrVoxevXrh7OzMBx98wPTp09m7dy9WVlaGip2FdNQUQgghdKSEjpoHDx4kMjKS2rVrayoUL2rbti3W1tbcu3ePkJAQvY6VF2mpEEIIIXSkhEdKw8LCAHB2ds52uUqlwsHBgaioKKKiooyaRVoqhBBCiELM09MTgIsXL5Kamppl+c2bNzWVidKlSxs1i1QqhBBCCB2lG+ilj+bNm2NnZ0doaCgLFizQqlhERETw4YcfAlC/fn08PDz0PFru5PaHEEIIoSMlPP3h4eHBhx9+yLRp01i9ejV79uyhevXqJCYmcu7cOWJiYvDw8GDGjBlGzyKVCiGEEEJHSpn6vGfPnvj4+LB27VpOnz7NoUOHsLKyokyZMvTs2ZN33nmH4sWLGz3HS1ep6NC+FaNHDaFOHT9sbW25cvUGa9euZ/mKNUbvcFO8jCe+TWtSvlYVyteuQqmqZbBSWfHbgvXsWbo5222cPV2o1qwW5WtVpnytKpSuVh5rW2v++fl3fpr0lVHzvqhtx1a0bNMU/9o1KFHSE1d3VxISErh25Sbbt+zm+1XrSUnJei/PVBo0rEuHTq1p2DiQqlUrYe9gR8STp5w4fpaVX//A338dM1u2TOa89nKj9HOr5HxKvu5KlCnBT0e/z9e6H/QYS9DR80ZOlLfRk99n6JhBACyas4IVn600c6LCpVatWnz22WdmzfBSVSomjB/G7FlTALhx4zaxcfHU8q/OF59/QuvWzejeY5BRv9xbDOxIy4EdC7RN3c5N6DGtv5ES5d+7w/tTr2EAiYlJPAp7zKULV/Aq6Ulg/doE1q9Ntzc789+ug4mOjjF5tmavNmLLb2sBSEtL49bNO8THJVCpcnk6d2lH5y7tWDB/GXM/+cLk2TKZ+9rLjZLPrZLzKf26S05K5vzxCzkuL+7ljncFb5ISk7hx8YYJk2WvctUKvDOsr7ljFJgSnv5QkpemUtGwQV0+mTmJtLQ0+v1vBOvXbwXA3786O7f/yOud2/HB6CF8tsh4f/3HRsRwfv8p7py7zp2gGzTu1Yo6HRvmuk1iTDyX/jzHnXM3uHPuOr5NatJiQAejZczJuu838+msJZw8dlarE1CdQH9WrF6If50ajP9wJP83YZbJs1lYWHDzxm1WLF3N5k07iIqMBjJm7psweQQfjHuPcROGcfrkOfbuPmjyfEq49nKj5HOr5HxKv+6ePn7K6G5jclw+efFEvCt488/eI8TFxJswWfZmLJxCSmoqp46dpVHz+uaOk29Kuf2hFC/N0x9TJo/C0tKSlavWab7UAYKCghk34WMAJowfjkplvHrWnqWb+eqd+exesplLh86RFJ+Y5zZHfznIsn6z2b5wPef3nyI+KtZo+XKzcd1Wjh4+meVxpTMng5j54acAtOvY0hzROH3qHI0CO7B65TrNFztASkoKs2Z8xr69hwDo2/9Ns+RTwrWXGyWfW1BuPqVfd7mxc7CjSbvGAOzfpPtIjobS8+0u1GsUwLIF3/Lg/kNzxxF6eCkqFc7OTrRunTEvx+rV67Is37hxO1FR0Xh4uNOyRWNTxyv0rl+9BYC9g71Zjh8bE0daWlqOyw/9cRiAylUqmiqSRmG/9sx9bvNiznxKvu7y0qxDU+wd7XkaHsnxgyfMmsWtuCvj/m8E1y7fYM2XP5o1iy7UBvpfUWHwSoVarebu3bsEBQUREaGMyWbq1M7oGJeQkMDpM1k7I6WmpnLy5DkA6tcPMHW8Qq9u/VoAXDinzAnLbG1tAUhMyLtlyNAK+7Wn9HOr5HzmvO7y0qZbxoyYf2w7SHqaoebZ1M2UmWNwc3dl+sR5pKbmXElTqnS12iCvoqLA7a3R0dF8++23BAUF4ebmRo8ePWjSpAkAhw8fZvr06dy7d0+zfv369ZkxYwbly5c3XOoCqvLsL4W7Ifdz/Mvi5q07tKaZZl2RO0tLS7xKevJa+xZM+mg0cbHxzJ1pvo6QuenyRsYslMeOnTb5sQvjtaf0c6v0fJnMed3lxt3LnTpNawOwf9N+s2Zp1KweXXp2ZOuGnZz4R1nlJHRToEpFbGwsb731Frdu3dL0eN29ezfz5s2jcuXKDB06lKSkJKytrTXjjB87doy3336brVu34u5unumZ3dxcAIh8mvOY55GRUVrriuwNeu9tPpo9Ueu93dsPsGDOUq5eum6mVDnr979e+NeuQVJSMl8tX2Py4xema0/p51bp+V5k7usuN627tsLKyoq710O4cu6q2XLY2Nrw8YLJREfFMHf652bLoa+i08ZgGAW6/bF69Wpu3rxJtWrVWLhwIQsXLqRatWosWLCAr776ChsbGxYtWsTZs2c5duwYO3fuJCAggPDwcFatWmWs3yFPdnZ2ACQnJ+e4TlJSxjL7Z+uK7IU9eMSJo6c5cyqIRw/DgWd/bXTrgKWlsrro+Neqzqx5UwGYM3MRt28Zd3a+7BSma0/p51bp+TIp4brLTeatj31mbqUYOmYQFSqVY9Hs5Tx5rIxb5bpQwiylSlKglor9+/fj7OzMmjVrKFasGABNmjShTZs27Nu3jxkzZtChw/PHHStVqsTSpUtp164dhw4dYty4cYZNn0+JiRn3NG1sbHJcx9Y2Y1lCovLufyrJjq172bF1r+bn2nVrMuezaYwYOwRXNxemjvvEjOmeK1e+DD9t+Ap7ezs2btjG0sXmGUSnMF17Sj+3Ss8HyrnuclLxlQpUqVGZ9PR09m8231MfmWNSXDh3iZ9WbzRbDkMoShUCQyhQ9f7u3bsEBARoKhQArq6u1K1bF4AWLVpk2cbd3R1/f3+tfham9vRZ07NrLs3Lrq4uWuuK/Dl76jz/6zWUxMQk/tu/B6XLlDJ3JLy8PNi4dTUlS5Vg7+4/GP7eJLNlKczXnhLP7YuUlk9J111OXuveBoCgY+d5FPrIbDk+mj8JK5UV08fPlcGjipgCtVSkpqZib5/10a3MJt7slmW+n910rKZy/XrGY2flynpjZWWVbYe5ShXLa60r8u9h2GOCL1wmILAW1f18Cb33wGxZXN1c2Lh1NZUqlefwX8cY2G+kXHt6UNK5zY5S8intusuOhYUFLbtkjOdh7rEpqtf0Ra1Ws+KHhVmWOTs7ATB4RD/eHtSTsNCHdG9r/lGFcyKVIm0Faqnw9PTk8uXLWd7PfC8oKCjLsrS0NIKDgzXzvZvDmbMXSE5Oxt7enoA6NbMsV6lUBAZmPJp2/Lj0QNaFyiqjfmqlsjJbBkdHB37e+A3Va/hy+lQQfXq9R2JiktnyQNG49pRwbnNj7nxKvO6yU7txLby8PUlKTOLPHX+aOw4qlQpPL48sLzv7jD9SnZwc8fTywK24m5mT5k76VGgrUKWiQYMG3Llzh6VLl5KamkpqaipLlizh7t27+Pv7M3v2bB4+fD4aWnp6OgsWLODBgwfUqVPH4OHzKyYmlgMH/gJgwIDeWZb36PEfXFyKER4ewcFDR0wdr9ArU9aban4+AFy6cMUsGWxsrPl+3QoC69XmUvBV3uw2iNjYOLNkeVFhv/aUcG5zY+58Sr3uspN560MJw3IHVmmJj2dgtq/NP/8GZEwo5uMZSKu6r5s1qyiYAlUqhgwZgq2tLcuWLaNOnToEBASwfPlySpYsyeLFi3n06BHt27dn4MCBjBw5krZt27JmzRosLS3p39+8zVdz5i4mPT2dQQN706tXF837/v7VWTD/IwAWLFxOSkqKuSIqVs1a1RkzaSjlypfJsuzV1k1Y+8sKrK2tObD3T+7cNn3fGUtLS75Z/TnNWzTi5s079OgyINdHOE1Nydee0s+tkvMp/bp7kY2dDU07ZIwnZO5bH0WNjKipzUJdwBtCJ0+eZPr06Vy/nvFceO3atfnkk0+oUqUKR48eZdSoUURFPf9g2dnZMXXqVHr27Kl3WJVNab22nzxpJDNnZDznnjlTpF8NX6ysrNixYz9duw8gPV330eXe826a6/JKdX0Z8s3zJ2BsHe2wtrUhKT6RlMTnjxzO7TSJyAdPAHAtVZxJO+ZqltnY22Jjb0tKUjJJcc+fFvh68AJunsr9L7Vt0bqNOtiwSSAbflsNwKOwxzy4/xBrG2tKlymFi2tGp92zp87Tv9dQnkZE6nSM+FTdm4u79ejE16sWAXDj+i0e5/B42sOwRwzqP0qnY0Qm6vfXp7GvPW8n3caAMcW51YeSrz1TXHcAtYpV0HnbTC27tODDZVN4Gh7Jm3XfMtgomiFJhn8UdO6Sj+j2VmeDTH1+9fFJA6XKWWCpZgbZz8kHfxlkP+ZW4BE1AwMD2b59O5GRkVhbW+Po6KhZ1rBhQ/bt28fBgwcJCwvD09OT5s2bU7x4cYOG1tWcuYsJCgpm1MjBBATUpGRJL85fuMzatetZtny1Xl/q+WFlbYWTe7Es79s62GHr8HyMAksrS63/zm4ba1sbrG2fP6ZoZW28+8mXLlzlo0lzaNK8IT6vVKZy1YpY21gTGRHJqRPn2P7rHrZs2J7rPAjG9OLjmpWrVMxxroW7d8z3BJK5r72cKP3cKjlfYbjuMmXe+lDCsNyiaCtwS4U56dtSYWx5tVSYm64tFaagT0uFKejbUmFsurZUCOVfe4ZoqTAWY7RUGJIpWioCShnme//0g78Nsh9zM89cy0IIIUQRUIj+LjcJ5YxtK4QQQohCTVoqhBBCCB0VpTEmDEEqFUIIIYSOitLjoIYglQohhBBCR+nSp0KL9KkQQgghhEFIS4UQQgihI7n9oU0qFUIIIYSO5PaHNrn9IYQQQgiDkJYKIYQQQkdy+0ObVCqEEEIIHcntD22FqlLhaueY90pm9OV9ZY/dbquyNneEHCWlypTz+rgfq9w5GJR83RUGhx5dNHeEHMWs6G3uCEJhClWlQgghhFASuf2hTSoVQgghhI7k9oc2efpDCCGEEAYhLRVCCCGEjuT2hzapVAghhBA6UqvTzR1BUaRSIYQQQuhIpj7XJn0qhBBCCGEQL1VLRYOGdenQqTUNGwdStWol7B3siHjylBPHz7Ly6x/4+69j5o5Ih/atGD1qCHXq+GFra8uVqzdYu3Y9y1esQW3GXsbly5ehZaumBAbWIrBubapVr4pKpeLjjxcwf95Ss+X6N6WWn9KzKTWf0q87pefLZM5z+/u1Bxy++YjgsCgexyYSmZCMnbUVlYo70+4Vb96sUwFrK+2/by8/jOL3qw84FfKEG09iiE1KxdnOmuolXOheqzytfEoZNXNBKOGzqyQvTaWi2auN2PLbWgDS0tK4dfMO8XEJVKpcns5d2tG5SzsWzF/G3E++MFvGCeOHMXvWFABu3LhNbFw8tfyr88Xnn9C6dTO69xhktgt42LCBDBs+0CzHzi8ll5+Ssyk5n9KvO6XnA/Of2++O3+Bs6FNsrCzxdLLDx6sY4bFJBN1/StD9p+y4eI8vezWimF3GIGkhT+N4a+2fmu1LuzjgXcyBe1HxHL71mMO3HtPZrwwfd6iNpYWF0XLnl9z+0PbS3P6wsLDg5o3bjP/gI3wqNqBh3fa0at6VqhUasGjBlwCMmzCMtu1bmCVfwwZ1+WTmJNLS0ujTdyi+1ZpQN/A16jVoT1jYI17v3I4PRg8xSzaA8CcR7Ny5nxkzFvJGl/5s2bLTbFmyo+TyU3I2pedT+nWn9HxKOLdd/cvzzVuNODy6Azvebc1P/Zqzd+hrfPd2U0o42xH8MIqlf13WrK8GPB1tGfVqNfYNfS1jm/7NOTiiHZPa+GEB/HbhHhvO3DZqbqGbl6ZScfrUORoFdmD1ynVERUZr3k9JSWHWjM/Yt/cQAH37v2mWfFMmj8LS0pKVq9axfv1WzftBQcGMm/AxABPGD0elMk/j0vx5S+nZ4x3mzV3Cvn2HiIuNN0uOnCi5/JScTen5lH7dKT2fEs5tl5plqVfOI8stDn9vN8a2rAHAH9fCNO+XcLZj25BWDGhQBU8nO837lhYWvBVQkR61ywOw+dxdo2UuCLVabZBXUfHSVCpiY+JIS0vLcfmhPw4DULlKRVNF0nB2dqJ166YArF69LsvyjRu3ExUVjYeHOy1bNDZ1PMVTcvkpORsoP5/QXWE4txWLOwGQmPL8u9lWZYW9dc6VnEYVPAG48zTWuOHyKV2tNsirqNC7UpGQkEB4eDgRERGkpBTeSaFsbW0BSExINPmx69TO6DyVkJDA6TPnsyxPTU3l5MlzANSvH2DqeIqn5PJTcjZQfj6hu8Jwbs+FPgWgWgmXfG+TlJoxLoStysoomYR+CtzmlZSUxObNm9m/fz/BwcFERkZqlllYWFCuXDkaN27Mm2++ySuvvGLIrEbV5Y32ABw7dtrkx67yrHXkbsj9HFtTbt66Q2uaadYVzym5/JScDZSfT+hOqec2LV1NeFwiB68/ZPGhS9hbWzHi1fz/W7H3yn0Aapd2N1bEApERNbUVqFJx+vRpRo8ezePHj7O9B6RWq7l9+za3b9/m559/5r///S+TJ0/GykrZNcp+/+uFf+0aJCUl89XyNSY/vptbRi098mlUjutERkZprSueU3L5KTnbi8dUaj6hO6Wd2x9O3mTB79rTuLesWpJhTX2p4lksX/v459YjTf+L/vUrGzyjLopSfwhDyHel4tatWwwcOJDExEQaNGhAw4YNcXV15d69e2zfvp0nT54wYcIEatasyeHDh/n555/58ccfefLkCYsWLTLm76AX/1rVmTVvKgBzZi7i9q0Qk2ews8vojJScnJzjOklJGcvs7exyXOdlpeTyU3I2UH4+oTulnVsvJztql3YjNV3Ng+gEnsQlceJuOLsuOTG0uDNWlrk/HvogOp6p288A8GadCtQtW9zomUXB5btSsXz5chITE5k2bRr//e9/tZaNHDmSd955h4ULF7Jz506GDx/OgAEDGDZsGLt376ZDhw60bdvW4OH1Va58GX7a8BX29nZs3LCNpYtXmiVHYmJGPw4bG5sc17G1zViWkGj6Ph9Kp+TyU3I2UH4+oTulndu2r3jT9hVvzc/n7z9l5t4gVh69TnRiClPb+ue4bVRCMsN+OcbThGQCyxZnXMvqRs+bXzJOhbZ8d9Q8cuQIvr6+WSoUkNHJccqUKSQlJbFzZ8Zz2o6OjixcuBBbW1s2bNhguMQG4uXlwcatqylZqgR7d//B8PcmmS3L02fNk665NEG6urporSueU3L5KTnbi8dUaj6hO6Wf25rebizt3gAbK0s2nbvD/ajsH8eNT05l+MZj3HwSS/USLnzRrT42CuqkKY+Uast3pSIqKory5cvnuDxz2b179zTvFS9enLp163Lx4sWcNjMLVzcXNm5dTaVK5Tn81zEG9htJamqq2fJcv34LgHJlvXPsf1KpYnmtdcVzSi4/JWd78ZhKzSd0VxjOrZezHb5exUhXw9VH0VmWJ6emMXrzCc4/iKRScSeW9WyAo62yBoKWR0q15btSUbx4ca5evZrj8suXM0ZEc3Jy0nrf1dWVuLg4HeMZnqOjAz9v/IbqNXw5fSqIPr3eIzExyayZzpy9QHJyMvb29gTUqZlluUqlIjCwFgDHj5v+6RSlU3L5KTkbKD+f0F1hObdp6Rn/oKb96x/W1PR0Jmw7xfG74ZRxdeDLNxvh5mBrjoiiAPJdqWjUqBF37tzhs88+y9JU8+TJEz7++GMsLCwIDAzUWvbo0SPc3NwMk1ZPNjbWfL9uBYH1anMp+CpvdhtEbKz5KzwxMbEcOPAXAAMG9M6yvEeP/+DiUozw8AgOHjpi6niKp+TyU3I2UH4+obvCcG5Do+K5+jijhcLnhSdA1Go103ae5eD1h3g62fHVm43wclZmR2G5/aEt35WK9957D3t7e7755hs6dOjAtGnTWLhwIWPGjOG1117jypUrVK9enZYtW2q2SUpKIjg4GF9fX6OELwhLS0u+Wf05zVs04ubNO/ToMiDXR61Mbc7cxaSnpzNoYG969eqied/fvzoL5n8EwIKFywv1AGPGpOTyU3K2wpBP6M7c5zY4LJLlf1/hXmTWP94O33zE8F+OkZqupmklL8q6OWqWzT9wkZ3BobjZ2/BVr4aUdnUwSj5DSEdtkFdRYaEuQBXp+PHjjBkzhvDw8IyNLSw0NaxatWqxePFiSpQooVn/2rVrrFq1ijZt2tC6dWu9w3oU89F52249OvH1qoxHW29cv8XjxxHZrvcw7BGD+o/S6RiRifq1ekyeNJKZMyYCz2cT9Kvhi5WVFTt27Kdr9wGkp6frvH9blbXO2zZsWJf1G77R/Ozk5IidnS1xcfEkvDAKaeNGnQgNfVDg/Sel6v+lZuzyK6rZjJ1PydedvkyRT9/PhjHPbcyKrC0gLzpxN5zBP2e0gng42uLlbEdqWsYjpTFJGb9XjZKuLO1RX3Nr41xoBP1/zJg2oaSzHSWL2ee4/zV9muZ6fPtBC/L9u+jKxckw42VExd4wyH7MrUA9XurXr8/+/fvZu3cvFy9eJCEhAbf/b+/eg6Ku/z2OP1E2Yb0tQUBARpqLhBQhpknmkFRqTmlFFx2zhC5oqd2OzuQkOZllF+1X4lgKds9RQ9ShFI1Lk3KTS4CmeUxFzOTSIhfXXeB7/vDsKj8WM1n2u8T7MeOMfD877Gt22d33fq4eHowaNYoxY9rvHT906FCWLVtmt7CdcfGyqiE33tDhGR/Hj52wed0Rlr39H375ZT/z5j5NeHgovr7elJb9ymefbWBVYrKqHzoajQYvr/Y72PXtq6Vv3wvfItTc6MyZHz9nzubM+Zz9787Z84G6z22Q9wD+Z3wIeceq+d/qeo7WNGBuVdC5abjZ35t7gvy4L8Qf114XOs1NLRfynKo3cqreuZcy/5uGLuzhH/VUqK0zPRWO0Nmeiq7WmW+MXc0ePRXCOTnz31134Myvjb/rqVCbI3oq+mnts8V5Q9O/Y3VVjzmlVAghhBBdy7kW/AohhBDdiBwo1pYUFUIIIcQV+jdtXGUPUlQIIYQQ/wJms5nk5GRSU1OpqKhAq9USERHB7Nmzuekmx5yXIkWFEEIIcYWcZa2D2WwmLi6OnJwcPD09iYqKoqqqivT0dDIzM1mzZg2RkZFdnkOKCiGEEOIKOcucirVr15KTk0NoaCjr16+3Hpmxfft2Xn75ZV555RV27dpF3759/+Y3dY6s/hBCCCGukDNs093c3Mz69esBWLx4cZszuCZPnsy4ceOora1l8+bNnbqfyyFFhRBCCNGNFRYWYjAYCAgIIDS0/eFxkyZNAmD37t1dnkWGP4QQQogr5AxzKiynhIeEhNhst0zStNyuK0lPhRBCCHGFFDv964yTJ08C4Ovra7Pdct1gMNDY2LU7P0tPhRBCCKGyvzt081JDF5ZCwd3d9uFrWq22zW27crJmtyoqqs8cUjuCEEIIYdVsqrTL7+nMSd6WIRgXFxe7ZOmMblVUCCGEEP9GnZlEael5aGpqstl+9uzZdrftKjKnQgghhOjG/Pz8ADh16pTNdst1nU4nRYUQQgghOhYcHAzA/v37bbaXl5cDEBQU1OVZpKgQQgghurHw8HB0Oh0VFRWUlZW1a//hhx+Azs3buFxSVAghhBDdmKurK0888QQAS5YsoaGhwdqWlpZGRkYGHh4ePPTQQ12exUVxhp07hBBCCHHFTCYTsbGx5OXl4enpyciRI6murqagoACNRkNiYiJ33nlnl+eQokIIIYT4FzCZTCQlJbF161br0ecjRoxgzpw5He62aW9SVAghhBDCLmROhRBCCCHsQooKIYQQQtiFFBVCCCGEsAspKoQQQghhFz3u7A+z2UxycjKpqanW2bERERHMnj3beua8GsrLy9mzZw+lpaWUlZVRWXn+kJrdu3cTEBCgWi44/5jl5uaSmZlJYWEhlZWVNDU14evryx133EFcXBz+/v6qZtywYQN79+7l4MGD1NTU0NjYyMCBAwkNDeWxxx4jKipK1XwXUxSFmTNnkpubC5xfRz5kyBDV8ixcuJCUlJQO2x999FGWLFniwES21dfXk5SUxK5duzhx4gQAPj4+jBgxgrlz5+Lj4+PwTLm5udb9AS5l1KhRfP755w5I1N6RI0f49NNPyc3N5fTp07i6ujJo0CCio6OZNWsW/fr1UyWXxdGjR1mzZg179uyhpqYGnU7H6NGjmT17NoMHD1Y1m/jnelRRYTabiYuLIycnB09PT6KioqiqqiI9PZ3MzEzWrFlDZGSkKtlWrVrVqQNlulJ+fj6xsbEA+Pv7ExERAcAvv/zC119/zdatW1m7di233nqrahmTk5OpqKhAr9cTHh6Om5sbFRUVZGZmkpmZyaxZs1iwYIFq+S62YcMGcnNzcXFxwZkWX91xxx1cc8017a6r+bxaHD58mKeeeorTp0/j5+fH2LFjaWlp4dixY2zatImpU6eqUlR4eXkxderUDtvT09NpaGjgtttuc2CqCwoKCoiNjcVoNBIYGMhdd93F2bNnKSwsZNWqVaSlpfHtt9+i0+lUyZeXl8ezzz5LU1MTgYGBREVFcfToUbZt20Z6ejrr1q2zvt+IbkLpQRITExW9Xq889NBDSn19vfX6tm3bFL1er4wePVppaGhQJduaNWuUFStWKOnp6cqpU6eUMWPGKHq9XqmoqFAlz8X27NmjzJ8/XykpKWlz3Wg0KgsXLlT0er0SFRWlmEwmlRIqSmFhoc3nLj8/XwkLC1P0er1SXFysQrK2/vjjDyU8PFyJjY1VoqKiFL1erxw+fFjVTAsWLFD0er2Sk5Ojao6O1NXVKWPHjlWGDRumfP7550pra2ub9mPHjik1NTUqpevYyZMnlWHDhilBQUHK8ePHVclw3333KXq9Xlm5cmWbx+2vv/5Spk6dquj1euXdd99VJVtTU5MSGRmp6PV6Zfny5W3yffnll4per1ciIyOVpqYmVfKJK9Nj5lQ0Nzezfv16ABYvXtymy2/y5MmMGzeO2tpaNm/erEq+Z555hvnz5xMdHa3KN65Luf3221mxYgU333xzm+t9+vQhISGB/v37U1lZSVFRkUoJz3+btnX6XkREBBMnTgRg7969jo7Vzuuvv05raysJCQlqR+k2Pv74Y/7880+efPJJZsyYgYuLS5v2QYMGcfXVV6uUrmOpqam0trYSERHBdddd5/D7/+uvv/jtt9/QaDTEx8e3edx0Oh2zZs0Czvc4qmHnzp1UVVURGBjISy+91Cbf9OnTGTVqFFVVVXz33Xeq5BNXpscUFYWFhRgMBgICAggNDW3XPmnSJKBzZ9r3RH369CEwMBCA06dPqxumA66u50f5rrrqKlVzbNmyhaysLObNm6f6PJnu4ty5c6SkpNCrVy+eeuopteP8I6mpqQBMmTJFlfvXaDSXdTsPD48uTmKb5eTMkSNH0rt373bto0ePBmDXrl0OzSU6p8fMqfj1118BOtyq1DJJ03I7cXlaWlqsk0q9vLxUTtPegQMH+P777+nduzdjx45VLUd1dTXLli1j+PDhzJgxQ7Ucl5Kenk56ejomk4lrr72WyMjIdr1TjlZWVsaZM2cICgrC29ub7Oxs9u7dS2NjIwEBAURHRzvlZL6SkhKOHDmCu7s7EyZMUCVDv379CAsLo7i4mNWrVzN37lxrb4DBYCApKQnAIYdM2XL27FkABg4caLPdMs+jo+O8hXPqMUXFyZMnAfD19bXZbrluMBhobGy02ZUu2ktNTaW2thZPT0/Cw8PVjsPmzZvJz8/HbDZTWVlJcXExrq6uJCQkMHToUNVyWU4OXLp0qc1vZc7giy++aPPzypUrGTduHMuXL1dtIt/hw4eB8xOEn3vuOTIyMtq0r1ixgmeffZb58+erkK5jW7ZsAeDuu+9WdXXFm2++SVxcHImJiaSlpREUFITRaGTfvn24u7vzzjvvOOSQKVssQ1aWLyX/zbLCR96Tu5ceM/zR2NgIgLu7u812rVbb7rbi0k6cOME777wDwIsvvqj68AKcH+ZKSUlh+/btFBUV4ebmxuLFi3n44YdVy7Rz50527NjBrFmzGDZsmGo5OjJs2DAWL17M999/T3FxMT/++CNvv/023t7eZGVl8dxzz9Ha2qpKtrq6OgCys7PJzs5m/vz5ZGVl8fPPP/P666+j0WhYvXo1GzduVCWfLSaTibS0NIBLrgxxhKFDh7JhwwZuueUWjh49yo4dO8jKyqKhoYGwsDBuvPFG1bJZVsRkZWVRVVXVps1oNLJ9+3brz/Ke3H30mKJC+f+le/89yUtcmfr6embPno3BYGDixInExMSoHQmApUuXcvDgQYqKitiyZQuTJk1i0aJFxMfHc+7cOYfnMRgMvPHGGwwaNIg5c+Y4/P4vx5NPPsm0adMYPHgw7u7u+Pv7M3XqVDZt2oROp6OoqIgffvhBlWwtLS3A+YnWM2fOJD4+Hl9fX7y8vJg+fTovvfQSAImJiarksyUjIwODwYCvr691XoBacnJyuP/++2loaGDt2rXk5+eTnZ3NkiVLyMnJYdq0afz000+qZBszZgy33HILTU1NxMbGsm/fPhobGzlw4ABPP/00NTU11tv26tVjPqq6vR7zTFm6zpqammy2W8b3Lr6tsM1oNBIfH8/BgwcZPXo0y5cvVztSO1qtluDgYN566y1iYmLIzMy0jiE70rJly6iurmbJkiW4ubk5/P47w8fHhwcffBBAtQ+ei1+LjzzySLt2y7WTJ09SUVHhsFyXYhn6eOCBB1T9MKyrq2PevHmYTCbWrl3L2LFjGTBgAD4+PtYNzc6dO0dCQoK1eHMkFxcXPvroI0JCQjh48CDTpk0jPDycKVOmUFxczKuvvmq93YABAxyeT1yZHjOnws/PD4BTp07ZbLdc1+l0UlRcgslk4oUXXiA/P59bb72VxMREpxj2uJQHHniAjRs3snv3buLj4x1637t376ZPnz4kJia2+zZt6fJdsGAB7u7uTJ8+XbVJfR1Re2XPxTu12tq1VavV4unpSU1NDVVVVaos3bxYbW2ttQBTa9WHRWZmJgaDgdtvv936/nexe+65B41Gw4kTJ6ioqLA+147k4+PDpk2byMjIoKCggMbGRvz9/Zk0aZL19XH99dc7/XuMuKDHFBXBwcFAxzOJLcubgoKCHJapu2lpaeGVV14hOzub4OBgPvnkk25RgFkmhNXW1qpy/+fOnSMvL6/D9tLSUgDGjx/vqEiXzTKnoaO5SF3t4q3zz5w5026FUWtrK2fOnAHazotSy/bt2zGbzYSFham+KsXyRal///42211dXdFqtdTV1VmfZzX06tWL8ePHt/v7tyzJVXsISfwzPaaoCA8PR6fTUVFRQVlZGcOHD2/TbhkzdsY3dmegKAqvvfYaO3bsYMiQISQlJXWbLsmcnBzg/DceRysoKOiw7a677qKyslL1sz86oigKO3fuBGj3enGUa6+9lpCQEMrLy8nJyWHy5Mlt2vft24fZbMbd3V31D3G4MPShdi8FYN1yvby8nObmZut+LRZHjhyxFhNqn93z3xoaGvj2229xcXHhscceUzuO+Ad6zJwKV1dX68E/luV9FmlpaWRkZODh4aHamm1n99Zbb5GSksJ1111HcnKyU+1gWFpaSkpKCiaTqV1bRkYGK1euBHCayaTOZP/+/Wzbtq3dY9fQ0MCiRYsoLS1Fq9Wq+rp45plnAPjggw84duyY9frp06dZunQpAA8//LDqXeS//fYb5eXlXHXVVdbN9NR055134ubmRmVlJe+99x7Nzc3WttraWhYtWgScX4Wh1h4zhw4dwmg0trlWU1PD888/T1VVFY8//ri1l1l0Dy6K4kQnGnUxk8lEbGwseXl5eHp6MnLkSKqrqykoKECj0ZCYmKjamu3MzMw2Y+779+/HbDYTHBxsfbMcN26cKisIdu3aZb3fUaNG2RyfBYiOjiY6OtqR0YAL+fr378/w4cPx9PSkvr6e33//nePHjwM41YFiFs7QU2F57AYOHMjw4cPx8PCgurqaAwcOUFdXh1arte5XoaaEhAS++eYbtFotYWFhuLq6UlRURH19PWFhYaxfv161IRqL5cuXs27dOiZMmMCHH36oahaLjRs3WreG9/Pz46abbsJoNFJSUkJ9fT1eXl58+eWX3HDDDarkW7hwITt37iQkJARvb28MBgMFBQUYjUbuvfde3n///cveGVQ4hx4z/AHnt2let24dSUlJbN26lR9//BGtVsv48eOZM2dOh7ttOkJtbS0lJSXtrh84cMD6f7W6dy1j1oD1uG5b/P39VSkqQkNDmTNnDvn5+fz+++/s27ePXr164e3tzf3338+jjz4qJx12ICgoiBkzZlBaWsqhQ4cwGAxoNBr8/f2ZMmUKTzzxhFNsKZ6QkMCIESP46quvKCkpobm5mcDAQCZPnszMmTPp06ePqvlaWlrYtm0boP7eFBeLiYlBr9fz2WefUVhYSFZWFr179yYgIICYmBji4uLw9PRULV90dDTV1dXWZeB9+/ZlxIgRxMTEWM/sEd1Lj+qpEEIIIUTX6TFzKoQQQgjRtaSoEEIIIYRdSFEhhBBCCLuQokIIIYQQdiFFhRBCCCHsQooKIYQQQtiFFBVCCCGEsAspKoQQQghhF1JUCCGEEMIupKgQQgghhF1IUSGEEEIIu5CiQgghhBB2IUWFEEIIIezi/wAqYSkcsYFauwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df_cm = pd.DataFrame(cmx)\n",
    "# plt.figure(figsize=(10,7))\n",
    "sns.set(font_scale=1.4) # for label size\n",
    "sns.heatmap(df_cm, annot=True, annot_kws={\"size\": 16}) # font size\n",
    "title = \"Confusion Matrix for SVM results\"\n",
    "plt.title(title)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Hand-Written Classification with SVM\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Create and fit the SVM model\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "svm_classifier = svm.SVC(gamma='scale')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "SVC(C=1.0, cache_size=200, class_weight=None, coef0=0.0,\n",
       "  decision_function_shape='ovr', degree=3, gamma='scale', kernel='rbf',\n",
       "  max_iter=-1, probability=False, random_state=None, shrinking=True,\n",
       "  tol=0.001, verbose=False)"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "svm_classifier.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Predict for our test set\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred_svm = svm_classifier.predict(X_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Get accuracy for the SVM model, we can see we have a nearly perfect model\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.9916666666666667\n"
     ]
    }
   ],
   "source": [
    "print(\"Accuracy: \"+str(accuracy_score(y_test, y_pred_svm)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's take a look at the confusion matrix for SVM, we can see a nearly perfect model with SVM\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "label_names = [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]\n",
    "cmx = confusion_matrix(y_test, y_pred_svm, labels=label_names)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df_cm = pd.DataFrame(cmx)\n",
    "# plt.figure(figsize=(10,7))\n",
    "sns.set(font_scale=1.4) # for label size\n",
    "sns.heatmap(df_cm, annot=True, annot_kws={\"size\": 16}) # font size\n",
    "title = \"Confusion Matrix for SVM results\"\n",
    "plt.title(title)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Comparing both SVM and Logistic Regression with K-Fold Cross Validation\n",
    "\n",
    "k-fold Cross validation is used when there are limited samples, the handwritten dataset contains about 1800 samples, this will give an opportunity for all the data to be in the training and test set at different given times. We will add <code>l2</code> regularization to visualize how well they both do against SVM.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/model_selection/_split.py:437: DeprecationWarning: `np.int` is a deprecated alias for the builtin `int`. To silence this warning, use `int` by itself. Doing this will not modify any behavior and is safe. When replacing `np.int`, you may wish to use e.g. `np.int64` or `np.int32` to specify the precision. If you wish to review your current use, check the release note link for additional information.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  fold_sizes = np.full(n_splits, n_samples // n_splits, dtype=np.int)\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/model_selection/_split.py:113: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  test_mask = np.zeros(_num_samples(X), dtype=np.bool)\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/model_selection/_split.py:113: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  test_mask = np.zeros(_num_samples(X), dtype=np.bool)\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/model_selection/_split.py:113: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  test_mask = np.zeros(_num_samples(X), dtype=np.bool)\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/model_selection/_split.py:113: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  test_mask = np.zeros(_num_samples(X), dtype=np.bool)\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/model_selection/_split.py:113: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  test_mask = np.zeros(_num_samples(X), dtype=np.bool)\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/model_selection/_split.py:113: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  test_mask = np.zeros(_num_samples(X), dtype=np.bool)\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/model_selection/_split.py:113: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  test_mask = np.zeros(_num_samples(X), dtype=np.bool)\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/model_selection/_split.py:113: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  test_mask = np.zeros(_num_samples(X), dtype=np.bool)\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/model_selection/_split.py:113: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  test_mask = np.zeros(_num_samples(X), dtype=np.bool)\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/model_selection/_split.py:113: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  test_mask = np.zeros(_num_samples(X), dtype=np.bool)\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/model_selection/_split.py:437: DeprecationWarning: `np.int` is a deprecated alias for the builtin `int`. To silence this warning, use `int` by itself. Doing this will not modify any behavior and is safe. When replacing `np.int`, you may wish to use e.g. `np.int64` or `np.int32` to specify the precision. If you wish to review your current use, check the release note link for additional information.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  fold_sizes = np.full(n_splits, n_samples // n_splits, dtype=np.int)\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/model_selection/_split.py:113: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  test_mask = np.zeros(_num_samples(X), dtype=np.bool)\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/utils/fixes.py:357: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.\n",
      "  if _joblib.__version__ >= LooseVersion('0.12'):\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/model_selection/_split.py:113: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  test_mask = np.zeros(_num_samples(X), dtype=np.bool)\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/utils/fixes.py:357: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.\n",
      "  if _joblib.__version__ >= LooseVersion('0.12'):\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/model_selection/_split.py:113: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  test_mask = np.zeros(_num_samples(X), dtype=np.bool)\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/utils/fixes.py:357: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.\n",
      "  if _joblib.__version__ >= LooseVersion('0.12'):\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/model_selection/_split.py:113: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  test_mask = np.zeros(_num_samples(X), dtype=np.bool)\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/utils/fixes.py:357: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.\n",
      "  if _joblib.__version__ >= LooseVersion('0.12'):\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/model_selection/_split.py:113: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  test_mask = np.zeros(_num_samples(X), dtype=np.bool)\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/utils/fixes.py:357: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.\n",
      "  if _joblib.__version__ >= LooseVersion('0.12'):\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/model_selection/_split.py:113: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  test_mask = np.zeros(_num_samples(X), dtype=np.bool)\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/utils/fixes.py:357: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.\n",
      "  if _joblib.__version__ >= LooseVersion('0.12'):\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/model_selection/_split.py:113: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  test_mask = np.zeros(_num_samples(X), dtype=np.bool)\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/utils/fixes.py:357: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.\n",
      "  if _joblib.__version__ >= LooseVersion('0.12'):\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/model_selection/_split.py:113: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  test_mask = np.zeros(_num_samples(X), dtype=np.bool)\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/utils/fixes.py:357: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.\n",
      "  if _joblib.__version__ >= LooseVersion('0.12'):\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/model_selection/_split.py:113: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  test_mask = np.zeros(_num_samples(X), dtype=np.bool)\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/utils/fixes.py:357: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.\n",
      "  if _joblib.__version__ >= LooseVersion('0.12'):\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/model_selection/_split.py:113: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  test_mask = np.zeros(_num_samples(X), dtype=np.bool)\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/utils/fixes.py:357: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.\n",
      "  if _joblib.__version__ >= LooseVersion('0.12'):\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/model_selection/_split.py:437: DeprecationWarning: `np.int` is a deprecated alias for the builtin `int`. To silence this warning, use `int` by itself. Doing this will not modify any behavior and is safe. When replacing `np.int`, you may wish to use e.g. `np.int64` or `np.int32` to specify the precision. If you wish to review your current use, check the release note link for additional information.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  fold_sizes = np.full(n_splits, n_samples // n_splits, dtype=np.int)\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/model_selection/_split.py:113: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  test_mask = np.zeros(_num_samples(X), dtype=np.bool)\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/utils/fixes.py:357: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.\n",
      "  if _joblib.__version__ >= LooseVersion('0.12'):\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/model_selection/_split.py:113: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  test_mask = np.zeros(_num_samples(X), dtype=np.bool)\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/utils/fixes.py:357: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.\n",
      "  if _joblib.__version__ >= LooseVersion('0.12'):\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/model_selection/_split.py:113: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  test_mask = np.zeros(_num_samples(X), dtype=np.bool)\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/utils/fixes.py:357: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.\n",
      "  if _joblib.__version__ >= LooseVersion('0.12'):\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/model_selection/_split.py:113: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  test_mask = np.zeros(_num_samples(X), dtype=np.bool)\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/utils/fixes.py:357: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.\n",
      "  if _joblib.__version__ >= LooseVersion('0.12'):\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/model_selection/_split.py:113: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  test_mask = np.zeros(_num_samples(X), dtype=np.bool)\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/utils/fixes.py:357: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.\n",
      "  if _joblib.__version__ >= LooseVersion('0.12'):\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/model_selection/_split.py:113: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  test_mask = np.zeros(_num_samples(X), dtype=np.bool)\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/utils/fixes.py:357: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.\n",
      "  if _joblib.__version__ >= LooseVersion('0.12'):\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/model_selection/_split.py:113: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  test_mask = np.zeros(_num_samples(X), dtype=np.bool)\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/utils/fixes.py:357: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.\n",
      "  if _joblib.__version__ >= LooseVersion('0.12'):\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/model_selection/_split.py:113: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  test_mask = np.zeros(_num_samples(X), dtype=np.bool)\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/utils/fixes.py:357: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.\n",
      "  if _joblib.__version__ >= LooseVersion('0.12'):\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/model_selection/_split.py:113: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  test_mask = np.zeros(_num_samples(X), dtype=np.bool)\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/utils/fixes.py:357: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.\n",
      "  if _joblib.__version__ >= LooseVersion('0.12'):\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/model_selection/_split.py:113: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.\n",
      "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n",
      "  test_mask = np.zeros(_num_samples(X), dtype=np.bool)\n",
      "/home/jupyterlab/conda/envs/python/lib/python3.7/site-packages/sklearn/utils/fixes.py:357: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.\n",
      "  if _joblib.__version__ >= LooseVersion('0.12'):\n"
     ]
    }
   ],
   "source": [
    "algorithm = []\n",
    "algorithm.append(('SVM', svm_classifier))\n",
    "algorithm.append(('Logistic_L1', logit))\n",
    "algorithm.append(('Logistic_L2', LogisticRegression(C=0.01, penalty='l2', solver='saga', tol=0.1, multi_class='multinomial')))\n",
    "\n",
    "\n",
    "results = []\n",
    "names = []\n",
    "y = digits.target\n",
    "for name, algo in algorithm:\n",
    "    k_fold = model_selection.KFold(n_splits=10, random_state=10)\n",
    "    if name == 'SVM':\n",
    "        X = flatten_digits\n",
    "        cv_results = model_selection.cross_val_score(algo, X, y, cv=k_fold, scoring='accuracy')\n",
    "    else:\n",
    "        scaler = StandardScaler()\n",
    "        X = scaler.fit_transform(flatten_digits)\n",
    "        cv_results = model_selection.cross_val_score(algo, X, y, cv=k_fold, scoring='accuracy')\n",
    "        \n",
    "    results.append(cv_results)\n",
    "    names.append(name)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We plot and we can see that SVM performs better all the time even with k-fold cross validation and it is better than both Logistic regressions on average\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = plt.figure()\n",
    "fig.suptitle('Compare Logistic and SVM results')\n",
    "ax = fig.add_subplot()\n",
    "plt.boxplot(results)\n",
    "plt.ylabel('Accuracy')\n",
    "ax.set_xticklabels(names)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## References\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "1.  [Recognizing Hand-written](https://scikit-learn.org/stable/auto_examples/classification/plot_digits_classification.html?utm_medium=Exinfluencer&utm_source=Nurture&utm_content=000026UJ&utm_term=10006555&utm_id=SkillsNetwork-Courses-IBMDeveloperSkillsNetwork-CV0101EN-Coursera-25797139&utm_email=Email&utm_campaign=PLACEHOLDER)\n",
    "2.  [MNIST classification using multinomial logistic + L1](https://scikit-learn.org/stable/auto_examples/linear_model/plot_sparse_logistic_regression_mnist.html?utm_medium=Exinfluencer&utm_source=Nurture&utm_content=000026UJ&utm_term=10006555&utm_id=SkillsNetwork-Courses-IBMDeveloperSkillsNetwork-CV0101EN-Coursera-25797139&utm_email=Email&utm_campaign=PLACEHOLDER)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h2>Authors</h2>\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "[Aije Egwaikhide](https://www.linkedin.com/in/aije-egwaikhide/?utm_medium=Exinfluencer&utm_source=Nurture&utm_content=000026UJ&utm_term=10006555&utm_id=SkillsNetwork-Courses-IBMDeveloperSkillsNetwork-CV0101EN-Coursera-25797139&utm_email=Email&utm_campaign=PLACEHOLDER) is a Data Scientist at IBM who holds a degree in Economics and Statistics from the University of Manitoba and a Post-grad in Business Analytics from St. Lawrence College, Kingston. She is currently pursuing her Masters in Management Analytics at Queens University. She is part of the IBM Developer Skills Network group where she brings her real-world experience to the courses she creates.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# References\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\\[1]  <a href='https://opencv.org/?utm_medium=Exinfluencer&utm_source=Exinfluencer&utm_content=000026UJ&utm_term=10006555&utm_id=NA-SkillsNetwork-Channel-SkillsNetworkCoursesIBMDeveloperSkillsNetworkCV0101ENCoursera25797139-2022-01-01'>Open CV</a>\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h2>Change Log</h2>\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<table>\n",
    "    <tr>\n",
    "        <th>Date (YYYY-MM-DD)</th>\n",
    "        <th>Version</th>\n",
    "        <th>Changed By</th>\n",
    "        <th>Change Description</th>\n",
    "    </tr>\n",
    "    <tr>\n",
    "        <td>2021-03-30</td>\n",
    "        <td>0.1</td>\n",
    "        <td>Aije</td>\n",
    "        <td>Created original version of the lab</td>\n",
    "    </tr>\n",
    "</table>\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Copyright © 2020 IBM Corporation. All rights reserved.\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python",
   "language": "python",
   "name": "conda-env-python-py"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}