More_Features.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import sklearn\n",
    "import sklearn.preprocessing\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import classification_report, plot_confusion_matrix\n",
    "from sklearn.decomposition import PCA\n",
    "import numpy as np\n",
    "\n",
    "import sklearn.pipeline\n",
    "from sklearn.compose import ColumnTransformer\n",
    "from transforms import FourierTransform"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [],
   "source": [
    "from dataset import HARDatasetCrops"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [],
   "source": [
    "dataset = HARDatasetCrops('motionsense-dataset/train', 256, 50, 50, metadata_file='motionsense-dataset/data_subjects_info.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dataset size: 4095\n",
      "Datapoints shape: (256, 12)\n"
     ]
    }
   ],
   "source": [
    "print('Dataset size:', len(dataset))\n",
    "\n",
    "sample = dataset[0][0]\n",
    "print('Datapoints shape:', sample.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Our dataset is composed of 5205 datapoints, each have shape `(256, 12)` because we have 12 signals of 256 samples each."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For our base line model we will use the **user-acceleration magnitude** and **rotation-rate magnitude**.\n",
    "The x,y,z components of the feature are stored on the last 3 columns, se we'll take those and get the magnitude/norm:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [],
   "source": [
    "X_acc = np.array([np.linalg.norm(sample[:,-3:], axis=1) for sample, _, _ in dataset])\n",
    "\n",
    "X_rot = np.array([np.linalg.norm(sample[:,-6:-3], axis=1) for sample, _, _ in dataset])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now we \"only\" have 256 samples per datapoint.\n",
    "\n",
    "Instead of directly using its value at each timestep, we will apply the Fourier transform and use the frequency domain as our input space."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The output of the Fourier transform is complex, we can take the norm to use real numbers. This will make it simetric so we can use only the positive frequencies."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [],
   "source": [
    "X_acc_fft = np.abs(np.fft.rfft(X_acc, axis=1))\n",
    "\n",
    "X_rot_fft = np.abs(np.fft.rfft(X_rot, axis=1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x12761edd0>"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(X_acc_fft[7], label='user-acceleration')\n",
    "plt.plot(X_rot_fft[7], label='user-rotation')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Classes"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The classes are the following:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'dws': 'down-stairs',\n",
       " 'ups': 'up-stairs',\n",
       " 'wlk': 'walk',\n",
       " 'std': 'standing',\n",
       " 'sit': 'sitting',\n",
       " 'jog': 'jogging'}"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset.CLASSES"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In order to fit a `sklearn` model we should encode each class with a unique integer:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [],
   "source": [
    "label_encoder = sklearn.preprocessing.LabelEncoder()\n",
    "label_encoder.fit(list(dataset.CLASSES.keys()));"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [],
   "source": [
    "y = label_encoder.transform([cls for _, cls, _ in dataset])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.countplot(label_encoder.inverse_transform(y));"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Train-test splitting"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We will first \"stack\" our features."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(4095, 512)"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X = np.hstack([X_acc, X_rot])\n",
    "X.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33, random_state=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of training examples: 2743\n",
      "Number of testing examples: 1352\n"
     ]
    }
   ],
   "source": [
    "print('Number of training examples:', y_train.shape[0])\n",
    "print('Number of testing examples:', y_test.shape[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### PCA feature reduction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [],
   "source": [
    "pca_acc = sklearn.pipeline.make_pipeline(\n",
    "    FourierTransform(),\n",
    "    PCA(n_components=32)\n",
    ")\n",
    "pca_rot = sklearn.pipeline.make_pipeline(\n",
    "    FourierTransform(),\n",
    "    PCA(n_components=32)\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "features = ColumnTransformer([\n",
    "    (\"acceleration\", pca_acc, slice(0, 256)),\n",
    "    (\"rotation\", pca_rot, slice(256, 512))\n",
    "])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Visualization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_feats = features.fit_transform(X_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(2743, 64)"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_feats.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.scatterplot(x=X_feats[:,0], y=X_feats[:,1], hue=label_encoder.inverse_transform(y_train));"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Classification"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [],
   "source": [
    "import sklearn.discriminant_analysis"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Pipeline(memory=None,\n",
       "         steps=[('columntransformer',\n",
       "                 ColumnTransformer(n_jobs=None, remainder='drop',\n",
       "                                   sparse_threshold=0.3,\n",
       "                                   transformer_weights=None,\n",
       "                                   transformers=[('acceleration',\n",
       "                                                  Pipeline(memory=None,\n",
       "                                                           steps=[('fouriertransform',\n",
       "                                                                   FourierTransform(window=None)),\n",
       "                                                                  ('pca',\n",
       "                                                                   PCA(copy=True,\n",
       "                                                                       iterated_power='auto',\n",
       "                                                                       n_components=32,\n",
       "                                                                       random_state=None,\n",
       "                                                                       svd_solver='auto',\n",
       "                                                                       tol=...\n",
       "                                                           steps=[('fouriertransform',\n",
       "                                                                   FourierTransform(window=None)),\n",
       "                                                                  ('pca',\n",
       "                                                                   PCA(copy=True,\n",
       "                                                                       iterated_power='auto',\n",
       "                                                                       n_components=32,\n",
       "                                                                       random_state=None,\n",
       "                                                                       svd_solver='auto',\n",
       "                                                                       tol=0.0,\n",
       "                                                                       whiten=False))],\n",
       "                                                           verbose=False),\n",
       "                                                  slice(256, 512, None))],\n",
       "                                   verbose=False)),\n",
       "                ('quadraticdiscriminantanalysis',\n",
       "                 QuadraticDiscriminantAnalysis(priors=None, reg_param=0.0,\n",
       "                                               store_covariance=False,\n",
       "                                               tol=0.0001))],\n",
       "         verbose=False)"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf = sklearn.pipeline.make_pipeline(\n",
    "    features,\n",
    "    sklearn.discriminant_analysis.QuadraticDiscriminantAnalysis()\n",
    ")\n",
    "clf.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "         dws       1.00      0.97      0.98       236\n",
      "         jog       1.00      1.00      1.00       265\n",
      "         sit       0.77      0.89      0.83       663\n",
      "         std       0.87      0.71      0.78       615\n",
      "         ups       0.90      1.00      0.94       298\n",
      "         wlk       1.00      0.98      0.99       666\n",
      "\n",
      "    accuracy                           0.90      2743\n",
      "   macro avg       0.92      0.92      0.92      2743\n",
      "weighted avg       0.90      0.90      0.90      2743\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_train, clf.predict(X_train), target_names=label_encoder.classes_))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<sklearn.metrics._plot.confusion_matrix.ConfusionMatrixDisplay at 0x128071dd0>"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAUMAAAEKCAYAAACIZDejAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAAgAElEQVR4nO3dd3hVVdbA4d9KQockQAKEXqRKTaKoFEVFARWsjFhxnLGggzoDjFgQsINlLDiKDvqJFQWFoYioIEgnoSMwgNQESChBQSC5Wd8f5yQkIeVibgPW63Mf7jln373XLa7sfco+oqoYY8zZLizYARhjTCiwZGiMMVgyNMYYwJKhMcYAlgyNMQawZGiMMYAlQ2PMaUZExonIXhFZU8R2EZHXRWSTiKwSkXhv6rVkaIw53XwA9Chme0+gqfu4B/i3N5VaMjTGnFZUdS6wv5gifYAP1bEIiBaRuJLqjfBVgMEiZSqolIsKdhi5OjSvG+wQzCnKDnYAhQilXsq2bVtJT0+X0tQRHtlANet3r8rq72lrgaN5Vo1V1bGn0FwdYEee5Z3uutTiXnT6J8NyUZRrfUeww8g1/6dRwQ7BnKLjWaGXDstGhE467NQxsdR1aNbvlGve16uyR1eMOaqqpW/0FJ32ydAYczoQkIAl+F1AvTzLdd11xQqdPz/GmDOXAGHh3j1Kbwpwh3tU+QIgQ1WLHSKD9QyNMYEipdrtmKca+RS4BIgRkZ3AU0AZAFV9G5gO9AI2AUeAu7yp15KhMSYAfDdMVtV+JWxX4IFTrdeSoTEmMHzUM/QXS4bGGP8TAnkA5Q+xZGiMCQCxnqExxgC+OlLsN5YMjTEBENDzDP8QS4bGGP8TbJhsjDGA9QyNMcaGySHiso7NeP6hPoSHCeOnLuFfH83Jt71ezWjeGHoTMdGVOfDrEe4d+RkpaRl07tCE5wZek1uuaf1Y7h7+CdPnrfVrvN8tWMfQl7/Ek53N7X0u4pH+V/i1PYunZD8sXMcT/5qEx5PNrb0vZOAd3fNtP3Y8kwdHfsSq9TuoGlWJsc/0p35cdZLXbmPQi58BoKoMvrsnvS5p5/d4Q+07Q4BwO4ACgIgMB35T1ZcC1SZAWJgw+u/Xcd0j75KyN4Mf3vsbM35ax4ate3PLjHzwaj77JpnPvkmiS3wTht3bg/ue+Zyflm+m613/AiC6SgWSP/8ns5ds9Gu8Hk82g0dN4Ks3H6R2zWguvXM0Pbu2oUXjEqdjs3j82OajL3/BhNceoHaNaK7880tc2aU1zRudaPOT/y4iukpFFn85jK9mJfH0mCm8+8xdtGgSx7fjBhEREc6e9Ay63fEiV3RuTUSE/xJDqH1nuUJ8n2Fo91t9IKFlPbbsTGdbyn4yszxM+m4lvTqfm69M84Y1mJe8CYB5yZvp2eXck+rp060t3y3awO/HMv0ab9LarTSuF0PDujGULRPB9d3jmf7jKr+2afEUL3ndNhrVjaVhHafNay+P55u5q/OV+Wbeavr2Oh+Aa7q156dlG1FVKpYvm5v4jh7PQvB/Qgi178zhDpO9eQSJX1sWkcdFZKOI/AQ0B8JEJMnd1k5EVETqu8ubRaSiiNwkImtEZKWIzC1tDHGxUezam5G7nJKWQVxsZL4yazelcvXFrQG4umtrIiuVp2pkxXxlrr+sHRO/W1HacEqUmpZBnZpVc5dr16xKalpGMa+wePwdz+60g9SuEX2izRrR7C7QphOXUyYiIpwqlcuzP+Mw4CSnrrc8xyW3Pc/oIX392is8EUvofGe5RLx7BInfkqGIJAA3A+1xZpA4D2dS4fIiEgl0AZYBXUSkAbBXVY8Aw4ArVbUd0LuIuu8RkWUiskwzvZs9tzhPvjmNTu0b8+O4h+jUoTG79h7Ek31iws+a1avQqnEtvl+8odRtmbNPwrkNmfvJY8wcN4jXPpzFUT+PLkJWiPcM/bnPsAvwlZvgEJEp7voFQCegK/Aczo1dBJjnbp8PfCAiE4BJhVXsTgE+FiCsci0tLojUtAzq1DhxW4DasVGkph3KV2b3vkPc8fh4ACpVKMs1F7fm0G8nZh2/9tK2TJ23liyP/2dEjouNYteeA7nLKXsOEBcbvNsaWDxQKzaalL0HT7S59yC1CrTpxHWQ2jWqkpXl4dffjlItqlK+Ms0a1qJSxXKs35JK+5b1/RZvqH1nQNB7fd4IRhqei5MoGwCTgXZAZ9xkqKr3AU/gzFSbJCLVS9NY8vqdNKkXQ/24qpSJCOf6y9sxY/66fGWqRVVE3C/qkdu78fG0Zfm233B5eybO8v8QGSC+VQM2b09j2650jmdmMWlWMj27tg1I2xZP4Tq0rM+WHWlsS9nH8cwsvv4umSu7tMlX5srOrZkwfQkA/529gs4JTRERtqXsIyvLA8CO1P1s2raHenHV/BpvqH1nuQI3uesf4s+e4VycHt7zbjvXAO/gJL1ngbmqmi0i+3GG0UMBRKSJqi4GFotIT5ykuO+PBuHxZDPklclMfOUvhIeF8fG0paz/ZQ9D776CFet3MmP+Ojp3aMKwe3uiKAtW/MLgV77KfX29WlWpUyOa+Su2/NEQTklERDijhvTlhoFj8HiUW3tfQMsmwTsKaPE4bT7/jxu5+eG38GRn0+/qC2jROI4Xx06jXcv69OjShluuuZAHR4yn440jiY6syDtP9wdgycrNvDH+OyIiwgkT4YVBfakeXdnv8YbSd+YI/fMMxZkH0U+VizwO3AnsBbYDyar6kojsAJ5W1bEi8hhws6q2dV8zCed+pwJ8DzysxQQZVrmWhtINoQ7YDaFOO3ZDqOJ16phIUtKyUo1xw6Lqa7nOg7wqe3T6Q0ln3A2hVPVZnF5gwfX18jx/DmffYc7y9f6MyRgTBDafoTHGwOkwTLZkaIwJDJvP0BhjCPlTaywZGmP8T2yYbIwxDusZGmMMuRc2hCpLhsYYv3Nm/bdkaIw524kgYZYMjTHGeobGGAOWDI0xBrBkaIwx7hGUYAdRvNM+GXZoXpf5ITRTTNXzHgx2CCc5sPTNYIcQ0kJphpgzlSDWMzTGGICwsND+o2PJ0BgTENYzNMYY22dojDGOUO8ZhvYg3hhzRsg5gOLNw6v6RHqIyAYR2SQijxayvb6IzBaR5SKySkR6lVSnJUNjTEBImHj1KLEekXBgDNATaAX0E5FWBYo9AUxQ1Q44929/q6R6LRkaY/xP8GXP8Hxgk6puUdXjwGdAnwJlFIh0n0cBKSVVavsMjTEBcQr7DGNEJO/Ny8eq6tg8y3WAHXmWdwIdC9QxHPhWRP4GVAIuL6lRS4bGmIA4hWSY7oNbhfYDPlDVl0XkQmC8iLRW1SLvC2vJ0Bjjdz6+AmUXUC/Pcl13XV53Az0AVHWhiJQHYnDu4V4o22dojAkM8fJRsqVAUxFpJCJlcQ6QTClQZjtwGYCItATKA2nFVWo9Q2OM/4nvLsdT1SwReRCYCYQD41R1rYiMBJap6hTgH8C7IvIIzsGU/qqqxdVrydAYExC+POlaVacD0wusG5bn+Tqg06nUacnQGBMYoX0Biu0zBPhuwTrOu2Ek8dcN59UPvg1qLG88eSsbZz7Pgs8eC2oceYXS5xOK8UDoxRRq8YBPzzP0i4AnQxFZEOg2i+PxZDN41AS+eG0AiyY8wcRvk1i/JTVo8Xw6dRE3DhwTtPYLCrXPJ9TiCcWYQi0e8D4RnlXJUFUvCnSbxUlau5XG9WJoWDeGsmUiuL57PNN/XBW0eBYs38yBQ0eC1n5Bofb5hFo8oRhTqMWTw5JhASLymzhGi8gaEVktIn9yt4WJyFsisl5EZonIdBG50Z/xpKZlUKdm1dzl2jWrkpqW4c8mTyuh9vmEWjwQejGFWjw5fHVtsr8E6wDK9UB7oB3OiZBLRWQuztGfhjgXX9cAfgbGFXyxiNwD3ANQr379wERsjCkVm8KrcJ2BT1XVo6p7gB+B89z1X6hqtqruBmYX9mJVHauqiaqaGBsTW6pA4mKj2LXnQO5yyp4DxMVGlarOM0mofT6hFg+EXkyhFg/g64ka/OKsP5oc36oBm7ensW1XOsczs5g0K5meXdsGO6yQEWqfT6jFE4oxhVo84F5cIt49giVYw+R5wL0i8n9ANaArMBgoB9zpro8FLgE+8WcgERHhjBrSlxsGjsHjUW7tfQEtm8T5s8livfdMfzolNKV6dGXWTH2aF8ZO56MpC4MWT6h9PqEWTyjGFGrxOEL/7nhSwhUqvm9Q5FececZG4UzOqMAzqvq5iIThTMJ4Cc4UPQK8qKqziqovISFR5y9eVtTmgLNbhZozTaeOiSQlLStVJitfq5k2uPMNr8puHNUjyQez1pyygPYMRaQ6sN+9RnCw+8ilqtkiMkhVf3PLLgFWBzJGY4wfBHkI7I2AJUMRqQ3MAV4qoehUEYkGygJPuwdSjDGnMQHCgnjajDcClgxVNQVo5kW5S/wfjTEm0KxnaIwxhP55hpYMjTH+Z/sMjTHGmfbfV5O7+oslQ2NMQFjP0BhjsH2Gxhhj+wyNMQZyrk0O7WxoydAYExAhngstGRpjAsOuQDHGGLFh8lknFGeIqXrdv4MdQj4b/+/PwQ4hn+qVywY7hJOEei/qVOXMZxjKLBkaYwIg9OcztGRojAmIEM+FlgyNMQEgoT/0t2RojPE7O8/QGGNclgyNMQbbZ2iMMYD1DI0xxiZqMMYYyJncNbSzoSVDY0xAhIV41zC05+E2xpwxRLx7eFeX9BCRDSKySUQeLaJMXxFZJyJrReSTkuq0nqExxu/EhxM1iEg4MAboDuwElorIFFVdl6dMU2Ao0ElVD4hIjZLqLTIZikhkcS9U1UPeBm+MMT7cZXg+sElVtwCIyGdAH2BdnjJ/Bcao6gEAVd1bUqXF9QzXAopz8niOnGUF6p9K9KHsuwXrGPryl3iys7m9z0U80v+KsyqeyzrU4/m7OxMeJoz/7mf+NWl5vu11Yyrz1sBLiapUlvCwMEaMX8Ss5O2UiQjj1fsupsM5sWRnK4/+Zz7z16b4NLa5S9bz7Jiv8WRnc1Ovjtzb77J825eu2syzYyazYUsqrz5xGz0ubueTdr9fuI6hr0wkOzub23pfyMN35v8Ojh3PZMCI8axcv4OqUZX4zzN3Ub92dfZnHOauR//D8p+3cfNVHRk1uG/ua57593/5fPoSMn49wvY5L/skzqKE2m8aTulyvBgRWZZneayqjs2zXAfYkWd5J9CxQB3NAERkPhAODFfVb4qNr6gNqlpPVeu7/9YrsOyXRCgi74lIK/f5Y/5ooyCPJ5vBoybwxWsDWDThCSZ+m8T6LamBaDok4gkLE0bf04Wbnp7KBQM/44bO59C8btV8Zf5xUwJfz9/Mxf/4krtfnsVL93YB4M7uLQHo9PAErhsxlWfuusinp094PNmMeH0S7z7/V6aPG8LUH5azaevufGXialTlhSE3c/VlHXza7pDRXzDhX/ez4LPHmVTId/DRlIVEV6nIsolPcf/N3RgxZjIA5cpGMPTeqxgx8LqT6r2yc2tmvT/IZ3EWF38o/abBvRzPy/+AdFVNzPMYW0L1hYkAmgKXAP2Ad0UkurgXeHUARURuzklOIlJXRBL+QHAlUtW/5Bn3ByQZJq3dSuN6MTSsG0PZMhFc3z2e6T+uCkTTIRFPQtMabEnNYNueX8nMymbST5vodX7D/IVUqVKxDACRlcqye/8RAJrXq8a81bsASM/4nYzDx+hwTom7Zry2av12GtSpTv3a1SlbJoKrunXguwVr85WpW6saLZrU9umRyuR122hUN4aGdZzv4LruCcyYuzpfmRlzV3PzVU5npPel7Zm7dCOqSqUK5bigfRPKlT150HVem0bUionyWZxFCbXfdI4w8e7hhV1AvTzLdd11ee0Epqhqpqr+AmzESY5Fx1dSqyLyJtANuN1ddQR426uQi6+3kohME5GVIrJGRP4kInNEJFFEXgAqiMgKEfm4tG0VJzUtgzo1T/SEatesSmpahj+bDKl44qpVYlf64dzllH2HiateKV+ZFz5fRt+Lm7Hm3duZ8MRVDHl3HgBrfkmnx/kNCQ8T6teoQvsmsdSpXtlnse1Jz6BW7Ik/5rVio9iT7v/vJnXvwfzfQY1oUtMO5i+TlkHtGk5sERHhRFauwP6Mw4SCUPtNAyDOfIbePLywFGgqIo1EpCxwMzClQJmvcXqFiEgMzrB5S3GVenM0+SJVjReR5QCqut8NoLR6ACmqehWAiEQB97ttPCoiD6pq+8JeKCL3APcA1Kt/xuy6DFk3dDmHT37YwJgpKzmveU3efvgyLnrocz76fj3N6lZl9ks3siPtV5as340nOzvY4ZoQ5avOu6pmiciDwEyc/YHjVHWtiIwElqnqFHfbFSKyDvAAg1V1X3H1epMMM0UkDOegCSJSHfDFL3418LKIvAhMVdV53h56d/chjAVISEjU0gQRFxvFrj0HcpdT9hwgLtb/Q5lQiSd1/2HqxJzoCdauXonUffl7OLdd1pKbRk4FYOmGPZQvE0H1yAqkZ/zO4+8vyC038/nr2Jziux5IzZgodufpke1Oy6BmAIaZcTWi838Hew8SF5t/d1NcbBQpbg8yK8vDod9+p1pUpYJVBUWo/abB2Wfoy10ZqjodmF5g3bA8zxX4u/vwijf7DMcAE4FYERkB/AS86G0DRVHVjUA8TlJ8RkSGlfASv4hv1YDN29PYtiud45lZTJqVTM+ubYMRSlDiSf7fXprERVO/RhXKRIRxfedzmLF0a74yu9J/o2vbugA0qxtNubLhpGf8ToWyEVQs5/w9vaRdXbI82WzYeaBgE39Ymxb12LornR2p+ziemcW02cu57KJzfVZ/UTq0rM+WHWlsS3G+g69mJdGza5t8ZXp0acNn0xYDMOWHFXRJbBYyExGE2m86R1iYePUIlhJ7hqr6oYgkAZe7q25S1TWlbVhEagP7VfUjETkI/KVAkUwRKaOqmaVtqzgREeGMGtKXGwaOweNRbu19AS2bxPmzyZCKx5OtDHl3HhOfuprwMOHj79ezfscBhvY7jxWb0pixdCtPvL+A1wZczIBr2qLAA6//AEBMVAUmPnU12aqk7jvMfa9979PYIsLDGfa367n7n2PxZCs39jyfpg1r8dr739C6eV0uu6g1q9Zv54GnPuDQb78ze+E6Xv+/mUwfN6R07UaE8+Kgm7hp4Ft4spVbrrmAFo3jeP6dabRvWZ+eXdtwW+8LuX/4hyTeMILoyIq898xdua9vf+1T/Hr4KJmZWUz/cTVfvj6AFo3jGP7G13w5M4kjRzNpffWT3N7nQv75116l/ZgKjT+UftNwaleXBIs4vckSCom0BTrjDJXnq2qpD02JyJXAaJwhdybO/sKXgEGquswdPvcGklX11qLqSUhI1PmLlxW12WB3xyuJ3R2veJ06JpKUtKxUAVVr1Eq7D/fuWOiE/vFJqppYmvb+iBJ7hiLyOHAL8BXO0P8TEflYVZ8vTcOqOhNnJ2del+TZ/k/gn6VpwxgTOkInvRfOmwModwAdVPUIgIg8CywHSpUMjTFnl1DZp1oUb5JhaoFyEe46Y4zxinM0OdhRFK+4iRpexdlHuB9YKyIz3eUrcE56NMYY78jpPblrzhHjtcC0POsX+S8cY8yZ6rQdJqvqfwIZiDHmzHVaD5NziEgT4FmgFVA+Z72qNvNjXMaYM0yo9wy9uQLlA+B9nOTeE5gAfO7HmIwxZyDx8hEs3iTDiu45gajqZlV9AicpGmOMV0QgPEy8egSLN6fWHHMnatgsIvfhzBtWxb9hGWPONKE+TPYmGT4CVAIG4uw7jAJC63oqY0zIC/Fc6NVEDYvdp79yYoJXY4zxmiAhf9/k4k66/gp3DsPCqOr1fonIGHPmOQ1mrSmuZ/hmwKIwfjV19E3BDiGfnq/MDXYI+SwZdnnJhUypnbb7DFXVt5PTGWPOWgKEn67J0BhjfOm0vwLFGGN84YxJhiJSTlWP+TMYY8yZyZn2P7SzoTf3TT5fRFYD/3OX24nIG36PzBhzRvHhTeT9E58XZV4Hrgb2AajqSpybyhtjjNdybgpV0iNYvBkmh6nqtgJdXI+f4jHGnIEEiAjxYbI3yXCHiJwPqIiEA38DNvo3LGPMmSbEc6FXyfB+nKFyfWAP8J27zhhjvCJyGl+Ol0NV9wI3ByAWY8wZLMRzoVczXb9LIdcoq+o9fonIGHNGOhPOM/wuz/PywHXADv+EY4w5EwkEdeJWb3gzTM43xb+IjAd+8ltExpgzT5DPIfTGH7kcrxFQ09eBBNN3C9Yx9OUv8WRnc3ufi3ik/xVnVTxJK/7H2A+/ITs7myu6xXNTny75tn81bQHfzk4mPCyMyMhKPHxvH2rERgMw7uNvWbb8f2Sr0qFNY+65s2eprzS48JzqDOrRnLAw4evkXfzfT1vzbf/7lc1IaFQNgPJlwqhWqSzdXpgDwOu3daBN3ShWbD/II5+sOKV2v1+4jqGvTCQ7O5vbel/Iw3fm/9yPHc9kwIjxrFy/g6pRlfjPM3dRv3Z1AF794Fs+/u9CwsLCeOEfN3LpBS0BeOezOXw4eQGqyh19LuK+fs4puqs37uQfL3zOseOZhIeHMXpIXxLObXiKn1TRQu03Dc6chqHMmytQDojIfvdxEJgFDPVF4yLysIhULGJbfxHx+zRiHk82g0dN4IvXBrBowhNM/DaJ9VtS/d1syMTjyc7m3+9PZ8Q/b+Wtlx7gxwVr2L5zb74yTRrG8eqz9/DmqAF07tiK9z+ZBcDPG7fz88YdvDHqfsaMHsDGLSms/nlrqeIJE/hnrxYM/Hg5N41ZwJWta9EotlK+Mq/M3Mitby/i1rcXMWHxDmb/fCLe8fO3MWzSmoLVlsjjyWbI6C+Y8K/7WfDZ40wq5HP/aMpCoqtUZNnEp7j/5m6MGDMZgPVbUvlqVhLzP32ML167n8GjJuDxZPPz5hQ+nLyAWe8PYu5HjzJz/hq27EgDYPgbkxnylx78+NGjDL3nKka8OfmUYy7uvYTSbxpO3Cr0tL0CRZw/8e2AWPdRVVUbq+oEH7X/MFBoMgyUpLVbaVwvhoZ1YyhbJoLru8cz/cdVZ008GzftIq5WNWrVrEaZiAi6XtiaRcs25CvT9txGlC9XFoDm59Qlff8hd4twPDOLrCwPmZlZeLI8VI2qXKp4zq0TxY79R9h14HeyPMq3a3ZzcfPYIstf0aYWM1fvzl1e+st+jhw/9WsCktdto1HdGBrWcT7367onMGPu6nxlZsxdzc1XdQSg96Xtmbt0I6rKjLmrua57AuXKlqFB7Rga1Y0hed02Nm7dQ8K5DahYviwREeF06tCUqXNWAs6R1V8PHwXg0G+/Uysm6pRjLkqo/aZznNbJUFUVmK6qHvdR5MzXJRGRSiIyTURWisgaEXkKqA3MFpHZbpm7RGSjiCwBOv3Rtk5FaloGdWpWzV2uXbMqqWkZgWg6JOLZd+AQsdUjc5djqkey78ChIst/OyeZhHbnANCyWT3atmrIHfe/xB33v0x8u3OoV6foxOWNGpHl2HPoxHwgew8do0ZkuULL1ooqT53oCiz9ZX+p2gRI3Xsw/+deI5rUtIP5y6RlULuGs3sgIiKcyMoV2J9xmNS0Ql679yAtGsexaMVm9mcc5sjR48xasJZdew4A8OwjN/DUG5Npc82TDHvja54c0LvU7yFvnKH0m84hIl49gsWbfYYrRKSDqi4vZVs9gBRVvQpARKKAu4BuqpouInHACCAByABmA4W2KSL3APcA1Ktfv5RhGW/NnreSTVtSeGHYXQCk7N7Hjl3pfDDm7wA88dx41qzfRusWDQISz5Wta/H9uj1k/+E/0f7VvFEtBt7RnRv/NoaKFcrSulldwsOc/sf7k37imYevp/el7fn6u2QGPvsxX735tyBH7D/OrUKDHUXxigxPRHISZQdgqYhsEJFkEVkuIsl/oK3VQHcReVFEuqhqwT9VHYE5qpqmqscp5kb1qjpWVRNVNTE2pnQ9kbjYqNy/1gApew4QF+u7IUuox1O9aiRp+070BNP3HaJ61ciTyq1YvZnPv57Hk4P6UaaM89NYuHQ9zZvWpUL5clQoX47EduewfmPpzrrae+gYNfP0BGtElmPvocJnjruidU1mrtld6LZTFVcjOv/nvvcgce5BotwysVGk7HV6i1lZHg799jvVoioRF1vIa90e5G29L+SHD4cw9Z2Hia5SgSb1nd/rZ9MWc023dgD0uawDyWu3++R95MQZSr/pHGHuVSglPbwhIj3cnLRJRB4tptwNIqIiklhifMVsW+L+2xtoDvQCbgJudP89Jaq6EYjHSYrPiMiwU63DH+JbNWDz9jS27UrneGYWk2Yl07Nr27MmnmZNapOyex+79x4gMyuLuQvX0DGheb4ym39J5c33pvLkoH5E59knGBsTxZqft+LxeMjK8rD6522lHiavSzlEveoVqR1dnohw4YrWtZi7Ie2kcg1iKlKlQhlW7fDN8K9Dy/ps2ZHGthTnc/9qVhI9u7bJV6ZHlzZ8Ns25WeSUH1bQJbEZIkLPrm34alYSx45nsi0lnS070ohv5fSO0/b/CsDO3fuZOmclN17p/D9ZKzaK+cmbAJi7bCNN6pXuc8sr1H7T4NsDKO4cCWOAnkAroJ+ItCqkXBXgIWBxwW2FKW6YLACqutmbikoiIrWB/ar6kXtU+i84tx+tAqTjBPyaiFQHDuEk3JW+aLs4ERHhjBrSlxsGjsHjUW7tfQEtm8T5u9mQiSc8PJz7+vdi2PPjyc5Wul/SgQb1avDRFz/QtFFtOia2YNwn33L06HFeeM05bhZbPYphg2+hU8dWrFr7Cw8M+TciEN/unJMS6anyZCujp2/gjdvjCRdhyvIUtqQd5t5uTfg55VBuYryydS2+LaRX+O5diTSMqUSFsuFM+3sXnp68jkWb95XYbkREOC8OuombBr6FJ1u55ZoLaNE4juffmUb7lvXp2bUNt/W+kPuHf0jiDSOIjqzIe884uwtaNI6jz+XxXHTzc4SHhzFq8E2Eu2PC/o++x/6MI5SJCGPU4L5EVXGOF/5raD8ee2UiWR4P5cqV4ZWhvrviNdR+0zl8uDvwfGCTqm5x6pXPgD7AugLlngZeBDCxpB4AABlVSURBVAZ7FV9Rx0REZCfwSlEvVNUitxVR35XAaCAbyMSZ7OFC4EGcfYndROQunNN2DgIrgOOq+mBx9SYkJOr8xctOJZSzzvxN6cEOIZ+HPirt7mffCsW744WF0BnKnTomkpS0rFQB1W/RVv/5nylelX2wc6NtOB2kHGNVdWzOgojcCPRQ1b+4y7cDHfPmChGJBx5X1RtEZA4wSFWLTRTF9QzDgcrgmzMlVXUmMLPA6mXAG3nKvA+874v2jDGhQzilnmG6qpa4j6/ItkTCcDpy/U/ldcUlw1RVHflHAzLGmFwCEb7r7e4C6uVZruuuy1EFaA3McU/VqQVMEZHexfUOS9xnaIwxpXWKPcOSLAWaikgjnCR4M3BLzkb3TJWY3LZ9MEy+rDTRGmNMXr6a3FVVs0TkQZzdbuHAOFVdKyIjgWWq6t3OyQKKTIaqWvrT+o0xxuXLi0tUdTowvcC6Qk/XU9VLvKnTbiJvjPE7wbtbcQaTJUNjjP+J74bJ/mLJ0Bjjd84VKJYMjTEm5E9PsWRojAmIEO8YWjI0xgRCcOcq9IYlQ2OM39nRZGOMcdkBFBN0nc6JKblQAC0b3j3YIeRT9bxiJ0YKigNL/X4vtMASbJhsjDE2TDbGGJf1DI0xBjvP0BhjECDceobGGGMnXRtjDCBIiA+ULRkaYwLCeobGmLOec2pNaGdDS4bGGP8T6xkaYwxgl+MZY4w7uWuwoyieJUNjTEDY0WRjjCH09xmG+rXTAfHdgnWcd8NI4q8bzqsffBvscCye0yyeN568lY0zn2fBZ4/5vO6S3uux45n8eeg44q8bzuX9R7M9ZV/utlfen0n8dcM574aRfL9wHQBHj2Vy2Z2j6XzL81zY9xmef2faSXX+86UvqNv17z5/L+Llf8Fy1idDjyebwaMm8MVrA1g04QkmfpvE+i2pFo/F47VPpy7ixoFjfF6vN+91/OSFREVWIPmr4dx/SzeGvzEZgPVbUpk0K5mFnz/Ol68PYNCLE/B4silXNoLJ/x7IT58MZe4nQ/l+4TqWrv4lt77l67Zx8NARn7+XnH2G3jyC5axPhklrt9K4XgwN68ZQtkwE13ePZ/qPqywei8drC5Zv5oAfEog373XG3FX0u6ojAH0u7cCPSzegqkz/cRXXd4+nXNkyNKgTQ+N6MSSt3YqIULliOQAyszxkZnlyZ5PxeLIZ9vrXjBh4rc/fCyKEefkIFr8nQxFpKCJr8iwPEpHhIjJHRF4TkRUiskZEzne3X+yuWyEiy0Wkij/jS03LoE7NqrnLtWtWJTUtw59NWjxnUDz+5M17Tdl7okxERDiRlSuwP+Pwya+tceK1Hk82XW55nmZXPMolHVuQ2LohAO9O+JGeXdtQKybKL+9HvHwES7APoFRU1fYi0hUYB7QGBgEPqOp8EakMHC34IhG5B7gHoF79+oGM15jTXnh4GPM+GUrGr0e4bfC7rNuUQtWoinz9/XKmvv2QX9o8He6bHOxh8qcAqjoXiBSRaGA+8IqIDASiVTWr4ItUdayqJqpqYmxMbKkCiIuNYteeA7nLKXsOEBfrn7+MFs+ZF48/efNea9c4USYry8Oh336nWlSlk1+79+TXRlWpSJeEZny/cB2rNuzklx1pxF8/gra9h3HkaCbx1w336fsJ9Z5hIJJhVoF2yud5rgXKqqq+APwFqADMF5EW/gwuvlUDNm9PY9uudI5nZjFpVjI9u7b1Z5MWzxkUjz958157dGnDp9MWAzD5h+V0Pa8ZIkLPrm2ZNCuZY8cz2bYrnc3b00g4tyHpB34l41dn/+bvR48ze8l6mjasyZWdW7Nh5vOsmjKSVVNGUrF8GZK/Gu7bNxTi2TAQw+Q9QA0RqQ78BlwNfONu+xMwW0Q6AxmqmiEiTVR1NbBaRM4DWgDr/RVcREQ4o4b05YaBY/B4lFt7X0DLJnH+as7iOcPiAXjvmf50SmhK9ejKrJn6NC+Mnc5HUxaWut6i3utzb0+lfcv69Lq4Lbf3uYj7nvqQ+OuGUzWyEv959i4AWjaJ49rLO3BB32eJCA9j9JC+hIeHsTv9EAOGj8eTnU12tnLd5fH06NKm1LF6I9SHyaJasHPmh0acIe9DwC5gC7AVuARYAVwMlAH+rKpLROQNoBuQDawF+qvqsaLqTkhI1PmLl/k1fnNms7vjFa9Tx0SSkpaVKpO1bNNBP5w8x6uy5zeJTlLVxNK090cE5ACKqr4OvJ53nYjMAT5S1YcLlP1bIGIyxgRYaHcMg3402RhzFnB2B4Z2NgxaMlTVS4LVtjEmwE6D+QyDfWqNMeYs4cuDySLSQ0Q2iMgmEXm0kO1/F5F1IrJKRL4XkQYl1WnJ0BgTAIKId48SaxIJB8YAPYFWQD8RaVWg2HIgUVXbAl8Co0qq15KhMSYgRLx7eOF8YJOqblHV48BnQJ+8BVR1tqrmXDC+CKhbUqWWDI0xfuftENnNhTEisizP454C1dUBduRZ3umuK8rdwIySYrSjycaYwPD+AEq6r84zFJHbgESc85mLZcnQGBMQPjy1ZhdQL89yXXdd/vZELgceBy4u7sKNHDZMNsYEhA/3GS4FmopIIxEpC9wMTMnflnQA3gF6q+pebyq1nqExxv98eJ6hqmaJyIPATCAcGKeqa0VkJLBMVacAo4HKwBfuEertqtq7uHotGRpjAsKXV6Co6nRgeoF1w/I8v/xU67RkaIzxOyH0r0CxZGiMCYgQz4WWDI1JW/R6yYUCrOpF/wh2CLmOrd/pm4pCPBtaMjTGBESoT+5qydAYExChnQotGRpjAiXEs6ElQ2OM39nkrsYYA6fF5K6WDI0xARHiudCSoTEmELybuDWYLBkaYwIixHOhJUNjjP+dyv1NgsWSoTEmMEI8G1oyNMYEhJ1aY4wx2D5DY4wBgTBLhoHx3YJ1DH35SzzZ2dze5yIe6X9Fvu3Hjmdy/1PjWbF+O9WiKjHuuT9Tv3Z1AF55fyYfTVlIeFgYLwy6kcsudG7B+uDIj5j50xpiqlZh4eeP59b1wthpfPj1AqpHVwbgyQd6c0WncwP2XgLtbIzn+4XrePzVSXiys7mt94U8dEf3fNuPHc/kgREfsXLDDqpFVuLdZ/pTv3Z15ixez9NvTSEzy0OZiHCG/+1auiQ2A+DZf09lwowlHPz1CNtmv1Sq+C7r2JznH76W8PAwxv93Mf8a/0O+7fVqVeWNx/5ETHQlDhw6wr0jPiElLQOAEQOupvtFLQkLE+Ys3cijr35dqli8F9rZMCTugSIiW0UkRkQaisiaU329x5PN4FET+OK1ASya8AQTv01i/ZbUfGXGT15IVGQFkr8azv23dGP4G5MBWL8llUmzkln4+eN8+foABr04AY8nG4B+V1/Al68/UGib9/frxrxPhjLvk6E+TYTevJdAOhvj8XiyefSlL/js1fuY/+ljfPVtEht+yd/Gx1MWER1ZkaVfDuO+fpcwcoxzC45q0ZX4+KV7mfvxUN4cdhsDRozPfc2VXc5l5rjST80VFiaMHnQ9N/3jXS64ZRQ3XN6B5g1r5isz8sFr+GzGMjrf8TKj3p/FsPt7AXB+64Z0bNuQzne8xEW3jaZDy3p06tCk1DGVJGdyVx/dA8UvQiIZllbS2q00rhdDw7oxlC0TwfXd45n+46p8ZWbMXUW/qzoC0OfSDvy4dAOqyvQfV3F993jKlS1DgzoxNK4XQ9LarQB0ij+HqpEVQ+69WDz+jSd53TYa1o2lYR2njWu7xzNj7up8ZWbMW82fep0PwDXd2jNv2UZUlbbN61ErNgqAFo3jOHosk2PHMwFIbN2IWjFRpY4voVV9tuzcx7aU/WRmeZj03XJ6dcn/B7l5w5rMS9oEwLykTfTs0hoARSlXNoKyEeGUKxNBmfBw0vb/WuqYvHEK900OioAkQxEZLCID3eevisgP7vNLReTjIl7TWESWi8h5JdWfmpZBnZpVc5dr16xKqjskyJGy90SZiIhwIitXYH/G4ZNfW+Pk1xbm3S/m0qnfczw48iMOHjpSYnlvefNeAulsjCc17SB1akSfaKNG9Elt7E7LoE5Np4zzeyrP/ozD+cr8d/YK2jarS7myZXwaX1xsFLv2HMxdTknLIC42f5JduymFqy9pA8DVF7chslJ5qkZWZOmabcxL3sz6/w5n/X+f4oclG9i4zaubx5Wa9Qwd84Au7vNEoLKIlHHXzS1YWESaAxOB/qq6tJDt94jIMhFZlpae5sewC/fnG7qw/KvhzPv4UWrGRPLEvyYFPAYT2tZvSeXpMVN46dE/BaX9J9/8L53aN+bHD/5Opw6N2bX3IJ7sbBrVqU7zhjU499qRtOozki4J53Bhu0YBiUlEvHoES6CSYRKQICKRwDFgIU5S7IKTKPOKBSYDt6rqysIqU9WxqpqoqomxMbHuX8oDudtT9hw46S9l7RonymRleTj02+9Ui6p08mv3nvzagmpUjyQ8PIywsDDuvLYTSWu3efMZeMWb9xJIZ2M8cbHR7Nqbp+e19+BJbdTK0ztzfk9HqRZVyS1/gDv/+R5vDrudRnVjfRob5PSO8/RcY6NO7rmmH+KOx/6Pi/u/wjPvzADg0G9HufriNixds43Dvx/n8O/H+W7hes5r3dDnMRbGhsmAqmYCvwD9gQU4CbAbcA7wc4HiGcB2oLO39ce3asDm7Wls25XO8cwsJs1KpmfXtvnK9OjShk+nLQZg8g/L6XpeM0SEnl3bMmlWMseOZ7JtVzqbt6eRcG7DYtvbnX7ihzd1zkpaNonzNlSfvJdAOhvj6dCyPr/sSGNbyj6OZ2bx9axkenRpk69Mjy6t+Xz6EsAZDndObIqIkPHrEW75+zs8OaA3Hds19mlcOZJ/3kGTujHUj6tGmYhwrr+8AzN+WpuvTLWoSrm9rEfuuIyPpzqx7txzkE4dmhAeHkZEeBidOjRh49Y9fokzL2+HyMEcJgfy1Jp5wCDgz8Bq4BUgSVW1QNf4OHAdMFNEflPVT0qqOCIinFFD+nLDwDF4PMqtvS+gZZM4nnt7Ku1b1qfXxW25vc9F3PfUh8RfN5yqkZX4z7N3AdCySRzXXt6BC/o+S0R4GKOH9CU83Pkbcffj7zM/6X/sO/gb5171BI/e04vb+1zEU69/zeqNOxER6sdV49XH+vnsQyrqvQTL2RhPREQ4zw+6kb4PvUV2djb9rr6AFo3jeGHsNNq3qE+Prm249ZoLGTBiPOfdOJKqkRUZ+3R/AN77Yh6/7EznpXHf8NK4bwD44rUBxFarwog3JjPx22X8fjSTttc8yW29L2TIX3udcnweTzZDXpnExFfvITxc+HjqEtb/soehf7mSFet3MuOntXSOb8Kw+3qhCgtWbGHwyxMBmDx7JV0TzmH++EGoKt8v3sA389f56qMrVqhfgSKqGpiGRC4DvgGiVfWwiGwE3lbVV0RkK+6+RGCqqrYWkWhgFvC0qk4pqt6EhESdv3hZAN6BOVNluadShZLYLoODHUKuY2s/Jvvw7lJlsvbxCTpr7mKvytaoUiZJVRNL094fEbCeoap+D5TJs9wsz/OG7tN0oLW77iBQ4pFkY8zpIbT7hWfQFSjGmFAmdqtQY4zJuQIllJ0RV6AYY0xpWc/QGBMQod4ztGRojAmIUD+1xpKhMcb/7L7JxhhzehxAsWRojAkIGyYbYwyh3zO0U2uMMQHhy1lrRKSHiGwQkU0i8mgh28uJyOfu9sUi0rCkOi0ZGmMCw0fZUETCgTFAT6AV0E9EWhUodjdwQFXPAV4FXiypXkuGxhi/EyBMxKuHF84HNqnqFlU9DnwG9ClQpg/wf+7zL4HLpISZY0/7fYbJyUnpFcqIL2ZXjcGZKCJUhFo8EHoxWTzF81U8DUpbQXJy0swKZSTGy+LlRSTvVFRjVXVsnuU6wI48yzuBjgXqyC2jqlkikgFUp5jP47RPhqrqk6mERWRZMKYNKkqoxQOhF5PFU7xQikdVewQ7hpLYMNkYc7rZBdTLs1zXXVdoGRGJAKKAfcVVasnQGHO6WQo0FZFGIlIWuBkoOAH0FOBO9/mNwA9awkzWp/0w2YfGllwkoEItHgi9mCye4oVaPD7h7gN8EJgJhAPjVHWtiIwElrkz4/8HGC8im4D9OAmzWAGb9t8YY0KZDZONMQZLhsYYA5ylyVBEhovIoGDHkUNEFgQ7hpKIyHs5Z/mLyGNBjuVhEalYxLb+IvJmoGMKRSKyVURiRKShiKwJdjyh7qxMhqFGVS8KdgwlUdW/qGrODXaDmgyBh4FCk6Exf9RZkwxF5HER2SgiPwHNgTARSXK3tRMRFZH67vJmEakoIjeJyBoRWSkic/0Y22/iGO22t1pE/uRuCxORt0RkvYjMEpHpInKjv2Jx26wkItPc971GRP4kInNEJFFEXgAqiMgKEfnYn3EUEctTQG1gtojMdsvc5X63S4BOfowlXw9LRAa5o4w5IvKa+5msEZHz3e0Xu+tWiMhyEani43gGi8hA9/mrIvKD+/zSor4bEWnsxmK34S3grDi1RkQScA6tt8d5z8lAEs5lP5FAF2AZ0MVNlntV9YiIDAOuVNVd7k3t/el6N752OJdRLXUTcCegIc4F6TWAn4Fxfo6lB5CiqlcBiEgUcD+Aqj4qIg+qans/x1BcLHcB3VQ1XUTigBFAApABzAaWByi2vCqqansR6Yrz/bQGBgEPqOp8EakMHPVxm/OAfwCvA4lAOREpg/N7zvnt5BKR5jjX8fZX1ZU+juW0d7b0DLsAX6nqEVU9xIkTNBfg/GC6As+5/3bB+ZEBzAc+EJG/4pzP5E+dgU9V1aOqe4AfgfPc9V+oaraq7sb5n93fVgPdReRFEemiqhkBaPOPxtIRmKOqae5F+58HPkQAPgVQ1blApPvHcz7witt7i1bVLB+3mQQkuH/QjwELcZJi3t9wjlhgMnCrJcLCnS3JsChzcX44DXB+KO1wks88AFW9D3gC57KeJBGpHqQ4A0pVNwLxOInoGbeHfNbHAmSR//+Z8nmeFzxhV1X1BeAvQAVgvoi08GUwqpoJ/AL0x/nDPg/oBpyDM4LIKwPYjvP7NoU4W5LhXOBaEang7re5xl0/D7gN+J+qZuOcqd4L+AlARJqo6mJVHQakkf96SF+bB/xJRMJFJBanl7oEp3dxg7vvsCZwiR9jAEBEagNHVPUjYDROMsor0x2O+V0RsfwK5Ox/WwxcLCLV3Zhu8mM4e4AablvlgKvzbMvZx9sZyFDVDPf3s1pVX8S5hMynydA1D2c4Ptd9fh+wvJBLz44D1wF3iMgtfojjtHdW7DNU1WQR+RxYCezF+WGiqltFRHB+SOAkwbqqesBdHi0iTXGmY/vefb1fQgS+Ai5021BgiKruFpGJwGXAOpwpiZJx/sr7Uxuc954NZOLsL3wpz/axwCoRSVbVW4MQy4XANyKSoqrdRGQ4zhDxILDCX4GoaqY4l3wtwZkIYH2ezUdFZDlQBvizu+5hEekGZANrgRl+CGse8DiwUFUPi8hRTh4i58R/WESuBmaJyG/uZWvGZZfjBZk79E5W1SLnjBORyqr6m1t2CdDJ3X9oQoCIzAEGqeqyksqa0HVW9AxDlTsEnEP+Xldhpro75MsCT1siNMb3rGdojDGcPQdQjDGmWJYMjTEGS4bGGANYMjzjiYgnzzWzX0gRs714WdclIjLVfd5bCrl5d56y0SIy4A+0UeiMQkWtL1Dmg1O5brvgtcbm7GbJ8Mz3u6q2V9XWOCfe3pd3ozhO+XegqlPcKyyKEg2ccjI0JlgsGZ5d5gHnuD2iDSLyIbAGqCciV4jIQhFJdnuQlQFEpIc4M+Yk40wmgbs+d95AEakpIl+JM7PMShG5CHgBaOL2Ske75QaLyFIRWSUiI/LUVXBGoWKJyF/delaKyMQCvd3LRWSZW9/VbvlwcWYEymn73tJ+kObMY8nwLCHO7RJ74lzjC9AUeEtVzwUO41yDfbmqxuPM4PN3ESkPvItz+WICUKuI6l8HflTVdjiXy60FHgU2u73SwSJyhdvm+Tiz8ySISFfJP6NQL5zJKUoySVXPc9v7Gbg7z7aGbhtXAW+77+FunEvkznPr/6uINPKiHXMWsZOuz3wVRCTnErV5OHcNqw1sU9VF7voLcKYIm+9cnUhZnMvbWgC/qOr/AETkI+CeQtq4FLgDQFU9QIaIVC1Q5gr3kTO9VmWc5FgFd0Yhtw1vLhFrLSLP4AzFK+PcJS3HBPc68/+JyBb3PVwBtM2zPzHKbXujF22Zs4QlwzPf7wXnHnQT3uG8q4BZqtqvQDlfzlkowPOq+k6BNh7+A3V9AFyrqitFpD/5J684afYYt+2/qWrepImINPwDbZszlA2TDcAioJOInAO5s0s3w5mIoKGINHHL9Svi9d/jTv7q7p+LIv/MMuD03v6cZ19kHRGpQdEzChWnCpDqzlJTcKKIm8SZ4acJ0BjY4LZ9v1seEWkmIpW8aMecRaxnaFDVNLeH9ak4U1MBPKGqG0XkHmCaiBzBGWYXNnX9Q8BYEbkb8AD3q+pCEZnvnroyw91v2BJY6PZMfwNuK2pGoRI8iTN1V5r7b96YtuNMZhEJ3KeqR0XkPZx9icniNJ4GXOvdp2POFnZtsjHGYMNkY4wBLBkaYwxgydAYYwBLhsYYA1gyNMYYwJKhMcYAlgyNMQaA/wdGmiOJvgmrQQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_confusion_matrix(clf, X_train, y_train,\n",
    "                      display_labels=label_encoder.classes_,\n",
    "                      cmap=plt.cm.Blues,\n",
    "                      normalize=\"true\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "         dws       0.96      0.89      0.92       123\n",
      "         jog       0.99      1.00      1.00       119\n",
      "         sit       0.78      0.84      0.81       341\n",
      "         std       0.79      0.67      0.72       285\n",
      "         ups       0.80      0.96      0.87       138\n",
      "         wlk       0.99      0.97      0.98       346\n",
      "\n",
      "    accuracy                           0.87      1352\n",
      "   macro avg       0.88      0.89      0.88      1352\n",
      "weighted avg       0.87      0.87      0.87      1352\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_test, clf.predict(X_test), target_names=label_encoder.classes_))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<sklearn.metrics._plot.confusion_matrix.ConfusionMatrixDisplay at 0x12c5a6e90>"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_confusion_matrix(clf, X_test, y_test,\n",
    "                      display_labels=label_encoder.classes_,\n",
    "                      cmap=plt.cm.Blues,\n",
    "                      normalize=\"true\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Metadata"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "metadata = np.array([metadata for _, _, metadata in dataset])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(4095, 4)"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "metadata.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = np.hstack([X, metadata])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33, random_state=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### PCA feature reduction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [],
   "source": [
    "pca_acc = sklearn.pipeline.make_pipeline(\n",
    "    FourierTransform(),\n",
    "    PCA(n_components=32)\n",
    ")\n",
    "pca_rot = sklearn.pipeline.make_pipeline(\n",
    "    FourierTransform(),\n",
    "    PCA(n_components=32)\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "features = ColumnTransformer([\n",
    "    (\"acceleration\", pca_acc, slice(0, 256)),\n",
    "    (\"rotation\", pca_rot, slice(256, 512)),\n",
    "    ('metadata', 'passthrough', slice(-4, X.shape[1]))\n",
    "])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Classification"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Pipeline(memory=None,\n",
       "         steps=[('columntransformer',\n",
       "                 ColumnTransformer(n_jobs=None, remainder='drop',\n",
       "                                   sparse_threshold=0.3,\n",
       "                                   transformer_weights=None,\n",
       "                                   transformers=[('acceleration',\n",
       "                                                  Pipeline(memory=None,\n",
       "                                                           steps=[('fouriertransform',\n",
       "                                                                   FourierTransform(window=None)),\n",
       "                                                                  ('pca',\n",
       "                                                                   PCA(copy=True,\n",
       "                                                                       iterated_power='auto',\n",
       "                                                                       n_components=32,\n",
       "                                                                       random_state=None,\n",
       "                                                                       svd_solver='auto',\n",
       "                                                                       tol=...\n",
       "                                                                   PCA(copy=True,\n",
       "                                                                       iterated_power='auto',\n",
       "                                                                       n_components=32,\n",
       "                                                                       random_state=None,\n",
       "                                                                       svd_solver='auto',\n",
       "                                                                       tol=0.0,\n",
       "                                                                       whiten=False))],\n",
       "                                                           verbose=False),\n",
       "                                                  slice(256, 512, None)),\n",
       "                                                 ('metadata', 'passthrough',\n",
       "                                                  slice(-4, 516, None))],\n",
       "                                   verbose=False)),\n",
       "                ('quadraticdiscriminantanalysis',\n",
       "                 QuadraticDiscriminantAnalysis(priors=None, reg_param=0.0,\n",
       "                                               store_covariance=False,\n",
       "                                               tol=0.0001))],\n",
       "         verbose=False)"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf = sklearn.pipeline.make_pipeline(\n",
    "    features,\n",
    "    sklearn.discriminant_analysis.QuadraticDiscriminantAnalysis()\n",
    ")\n",
    "clf.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "         dws       0.98      0.99      0.99       236\n",
      "         jog       1.00      1.00      1.00       265\n",
      "         sit       0.78      0.89      0.83       663\n",
      "         std       0.87      0.72      0.79       615\n",
      "         ups       0.93      1.00      0.96       298\n",
      "         wlk       1.00      0.98      0.99       666\n",
      "\n",
      "    accuracy                           0.91      2743\n",
      "   macro avg       0.93      0.93      0.93      2743\n",
      "weighted avg       0.91      0.91      0.90      2743\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_train, clf.predict(X_train), target_names=label_encoder.classes_))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<sklearn.metrics._plot.confusion_matrix.ConfusionMatrixDisplay at 0x1279520d0>"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_confusion_matrix(clf, X_train, y_train,\n",
    "                      display_labels=label_encoder.classes_,\n",
    "                      cmap=plt.cm.Blues,\n",
    "                      normalize=\"true\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "         dws       0.96      0.94      0.95       123\n",
      "         jog       1.00      1.00      1.00       119\n",
      "         sit       0.78      0.84      0.81       341\n",
      "         std       0.79      0.67      0.72       285\n",
      "         ups       0.84      0.96      0.90       138\n",
      "         wlk       0.99      0.98      0.98       346\n",
      "\n",
      "    accuracy                           0.87      1352\n",
      "   macro avg       0.89      0.90      0.89      1352\n",
      "weighted avg       0.87      0.87      0.87      1352\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_test, clf.predict(X_test), target_names=label_encoder.classes_))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<sklearn.metrics._plot.confusion_matrix.ConfusionMatrixDisplay at 0x117143550>"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAUMAAAEKCAYAAACIZDejAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAAgAElEQVR4nOzdeZxN9f/A8dd7ZuyZGcxgrIOsWccShagUJQmJtGj5qnxLGxWVbG1afFVa1K9v30qKqCSKZAuFGbsQsjQGYxs7M/e+f3+cO9PMYObOcu9cvJ897qN7zvmcz+d9rzPv+znb54iqYowxF7uggg7AGGMCgSVDY4zBkqExxgCWDI0xBrBkaIwxgCVDY4wBLBkaY84zIvKxiOwVkbXnWC4i8paIbBaR1SIS4029lgyNMeebT4COWSzvBNT0vPoB73lTqSVDY8x5RVUXAAeyKHIz8Kk6fgPCRSQqu3pD8ivAgiKFiqsUDS/oMNI0qVWhoEMwOeQOwJuwRAo6gn/s2L6Nffv25Smi4NCqqiknvCqrJxLXASfTzRqvquNz0FxFYGe66b898xKyWun8T4ZFwykS80BBh5Fm0ZwXCjoEk0MnT7sKOoQzhAQHTjZse0WLPNehKScoUrunV2VPrhx3UlWb5bnRHDrvk6Ex5nwgIH47KhcPVE43XckzL0t2zNAY43sCBAV798q7acBdnrPKLYEkVc1yFxmsZ2iM8Zd8OhAqIhOBdkCEiPwNvAAUAlDV94EZwA3AZuA4cI839VoyNMb4Qf7tJqtq72yWK/DvnNZrydAY4x+BdIr8LCwZGmN8T/DnCZRcsWRojPEDsZ6hMcYA+XWm2GcsGRpj/MCv1xnmiiVDY4zvCbabbIwxgPUMjTHmfNhNDuzo8sk1zWuw9JN/E/vpIzzW68ozllcuG8a3r93Jrx8+yPdv3E2FiJIZlpcsXpi1Xz7O6Ec6+SXenxevp3n3EcTcMowxn8zyS5sWT0a//PYHrXu9SKueI3n7s9lnLD91OoUHnv+EVj1HcsO/3mRnwv4My//efYAa1w7ivS9+SZv3+EtfUP/GZ2l3x8u5imnOkvW07DmK5j1GMPbTs8WUzP3P/pfmPUZw/b1vsGOXE9OBpGN07f8WVdsP5OnXJ2dY53RyCk+8/CWX3zqSVreN4vtfVuYqtmwJEBzs3auA+C0ZisgwERnor/ZSBQUJrw24gVsHT6DlvePofnV9aleNyFBmxIMd+HL2alr/631Gfzafofdfk2H5kHuuZsnq7X6J1+VyM2j0JCaP7c9vk55jyqxYNmzN9rZKiyef2xjyxmQmvPEA8ycM5tuf49j41+4MZSZOX0JYyWIsmfQ8/W5rx6h3v8+wfNjb33J1y3oZ5vW8oQVfvPlgrmN65vXJfDnmQRZNHMI3s2LZ+FfGzz1h2m+EhxZn2ddDebB3O0aMmwZAkcIhPNPvRoY/0vWMesd8MouIUpfw++TnWTRxCFfEXJqr+Lwi4t2rgFzwPcOmdSqyNf4A2xMOkZziZurcddxwRZ0MZWpXjWThir8AWLhyG53SLW9UM4qypUrwS+wWv8Qbu24b1StHEF0pgsKFQujWIYYZ81f7pW2Lx7Hij+1EV4qkakWnjZuvieGnhWsylPlx4Vp63uAMbdW5XSMWxm7CuQsMZi5YTZWoMtSuVj7DOq0aX0qp0OK5iiluvRNTtCemrh1imLkgY0wzF67hNk9MN7VvzMLlTkwlihWhZeMaFClc6Ix6v/j+Nx69uwMAQUFBlAm/JFfxZc+zm+zNq4D4tGUReVZENonIr0BtIEhEYj3LGomIikgVz/QWESkuIreKyFoRWSUiC/IaQ1RESeITD6dN70o8TFSm3eB1W/bQuU1dADq3rkNoiSKUCi2GCIx68Dqef99/u4YJiUlULFcqbbpCuVIkJCb5rX2LB3YnJlGx7D8DBkeVDWd3pjZ2Jx6iQlknjpCQYEJLFOVA0jGOHT/FuM/n8OS9WY1Kn3MJiYcyxFShbPgZn3t3YhIVy4X/E9MlTkznknTkOACvfPADV981mnuHfMze/YfPWT7PLtaeoYg0BXoBjXFGkGgOuIGiIhIKtAGWA21EpCqwV1WPA0OB61W1EdDlHHX3E5HlIrJck4/nOdbnP5jFlQ2rMv/9flzZKJr4xMO4XG7u79Kc2Uv/ZNe+I3luw1wcXv94Jv1ua0eJ4kUKOpRspbjc7Np7iOYNq/HLp0/RvH40w97+1ncNBnjP0Jdnk9sA33gSHCIyzTN/MXAl0BZ4CefBLgIs9CxfBHwiIpOAqWer2DME+HiAoJIVshy0PWHfESpGhqZNV4gMJSFTctu9/yh3DZsEQImihbipTV0OHztF83qVaNWgKvd1aU6JYoUpFBLMsROnGf7RHK+/hJyKigwjfs/BtOldew4SFRnms/YsnjOVjwwjfu+htOmEvYcon6mN8pHh7Np7kAplw0lJcXH42ElKh5Ugbt12ps9dxch3p3H46AmCRChSOIR7e7TNU0xRkeEZYtq199AZn7t8ZBjxe5wea0qKi8NHnZjOpXRYCYoXLUzndo0A6HJNEyZ8/1ue4jynAu71eaMg0vACnERZFfgOaAS0xpMMVfVB4DmckWpjRaRMXhqL2xBPjYplqFI+nEIhQXRrfxkzF2/MUKa0Z5cY4PHb2zDhxxUA9Hv5Gxrc/h8a9RnL8x/M4qvZq3yaCAFi6lVly45Etsfv43RyClNnx9GpbUOftmnxZNS4ThX++juRHbv2czo5he/mxHF96/oZylzfuj6TZiwFYPq8VbRuWhMR4bv3HmXZlBdYNuUF/tXzKgbc1SHPiRCgSd0q/LUzke2emL6dHUfHNg0ylOnYpj5feWL6fu5KWjdzYjoXEeG61vVZFLcZgAXLNlEr03HOfOW/wV1zxZc9wwU4PbyXPe3cBHyAk/ReBBaoqltEDuDsRg8GEJEaqvo78LuIdMJJivvP1oA3XG7lqbdnMOXVOwgOEibMXMmG7YkM7tuOlRt3MXPJJlo3jmbofdegwOLV2xn01ow8ffC8CAkJZvRTPek+YBwul9KnS0vq1sj2wV4WTz638dLj3en9xHu4XG56dW5J7epRjP5wBo3qVOb6Ng3o3bklj4z8nFY9RxIeWpz3h9+dbb0PvfA/Fq/YzIFDR4npOpSB93Xi9ptaeR3TywN70PPRd3G73fTu3JI61aN4ZfwPNK5ThY5tG9Dnplb0H/4ZzXuMoFRoccaP7Ju2fkzXYRw5fpLTySnMnL+ayW/1p3a1KIb+uwv9h3/Gc2OmUqbUJbz13O25/NayE/jXGUrqGTCfVC7yLHA3sBfYAcSp6usishMYqarjRWQI0EtVG3rWmYrzvFMB5gCPaRZBBpWsoIH0QKiD9kCo8449ECprba9oQVzs8jwFFBRWRYu09u7KupMzHo294B4Ipaov4vQCM8+vnO79SzjHDlOnu/kyJmNMAbDxDI0xBs6H3WRLhsYY/7DxDI0xhoC/tMaSoTHG98R2k40xxmE9Q2OMIcsLwAOBJUNjjM85o/5bMjTGXOxEkCBLhsYYYz1DY4wBS4bGGANYMjTGGM8ZlIIOImvnfTJsUqsCiwJopJhSzR8u6BDOcHDZOwUdQkArWjiwbxMraPmRwwSxnqExxoDzwKlAZsnQGOMX1jM0xhg7ZmiMMY5A7xkG9k68MeaCkHoCxZuXV/WJdBSRjSKyWUSeOcvyKiIyV0RWiMhqEbkhuzotGRpj/EKCxKtXtvWIBAPjgE5APaC3iNTLVOw5YJKqNsF5fvu72dVrydAY43tCfvYMWwCbVXWrqp4GvgRuzlRGgdQHpocBu7Kr1I4ZGmP8IgfHDCNEZHm66fGqOj7ddEVgZ7rpv4HLM9UxDJglIo8AJYBrs2vUkqExxi9ykAz35cOjQnsDn6jqGyLSCvhMROqrqvtcK1gyNMb4XD7fgRIPVE43XckzL737gI4AqrpERIoCETjPcD8rO2ZojPEP8fKVvWVATRGpJiKFcU6QTMtUZgdwDYCI1AWKAolZVWo9Q2OM70n+3Y6nqiki8jDwExAMfKyq60RkBLBcVacBTwIfisjjOCdT+qqqZlWvJUNjjF/k50XXqjoDmJFp3tB079cDV+akTkuGxhj/COwbUOyYIcDPi9fTvPsIYm4ZxphPZhVoLG8/34dNP73M4i+HFGgc6QXS9xOI8UDgxRRo8UC+XmfoE35PhiKy2N9tZsXlcjNo9CQmj+3Pb5OeY8qsWDZsTSiweCZO/40eA8YVWPuZBdr3E2jxBGJMgRYPeJ8IL6pkqKpX+LvNrMSu20b1yhFEV4qgcKEQunWIYcb81QUWz+IVWzh4+HiBtZ9ZoH0/gRZPIMYUaPGksmSYiYgcFcdrIrJWRNaIyG2eZUEi8q6IbBCR2SIyQ0R6+DKehMQkKpYrlTZdoVwpEhKTfNnkeSXQvp9AiwcCL6ZAiydVft2b7CsFdQKlG9AYaIRzIeQyEVmAc/YnGufm67LAH8DHmVcWkX5AP4DKVar4J2JjTJ7YEF5n1xqYqKouVd0DzAeae+ZPVlW3qu4G5p5tZVUdr6rNVLVZZERkngKJigwjfs/BtOldew4SFRmWpzovJIH2/QRaPBB4MQVaPEB+D9TgExf92eSYelXZsiOR7fH7OJ2cwtTZcXRq27CgwwoYgfb9BFo8gRhToMUDnptLxLtXQSmo3eSFwAMi8j+gNNAWGAQUAe72zI8E2gFf+DKQkJBgRj/Vk+4DxuFyKX26tKRujShfNpmlj0b15cqmNSkTfglrp4/klfEz+HzakgKLJ9C+n0CLJxBjCrR4HIH/dDzJ5g6V/G9Q5AjOOGOjcQZnVGCUqn4lIkE4gzC2wxmiR4BXVXX2uepr2rSZLvp9+bkW+509KtRcaK68vBmxscvzlMmKlq+lVe9+26uym0Z3jM2HUWtyzK89QxEpAxzw3CM4yPNKo6puERmoqkc9ZZcCa/wZozHGBwp4F9gbfkuGIlIBmAe8nk3R6SISDhQGRnpOpBhjzmMCBBXgZTPe8FsyVNVdQC0vyrXzfTTGGH+znqExxhD41xlaMjTG+J4dMzTGGGfY//wa3NVXLBkaY/zCeobGGIMdMzTGGDtmaIwxkHpvcmBnQ0uGxhi/CPBcaMnQGOMfdgeKMcaI7SZfdAJxhJjyd39e0CFk8Psb3Qo6hAwqly5W0CGcIdB7UTmVOp5hILNkaIzxg8Afz9CSoTHGLwI8F1oyNMb4gQT+rr8lQ2OMz9l1hsYY42HJ0BhjsGOGxhgDWM/QGGNsoAZjjIHUwV0DOxtaMjTG+EVQgHcNA3scbmPMBUPEu5d3dUlHEdkoIptF5JlzlOkpIutFZJ2IfJFdndYzNMb4nOTjQA0iEgyMAzoAfwPLRGSaqq5PV6YmMBi4UlUPikjZ7Oo9ZzIUkdCsVlTVw94Gb4wx+XjIsAWwWVW3AojIl8DNwPp0Zf4FjFPVgwCquje7SrPqGa4DFOfi8VSp0wpUyUn0geznxesZ/MbXuNxu7rz5Ch7ve91FFU/7BlGM7NOc4CBhwvzNvPPDugzLK5Yuzlv9riC0eGGCg4QXJ61gzupdGZYvePkmXv92Ne/N/CPP8SxavpHR73+H263c0rEF9/Zsn2F57JqtvPbBNP78azevPHM7Hdo0TFsWc+PTXBpdHoCoyHDGDrvH63bnLFnP4Den4Ha7uaNLKx67O+P3fup0Mv2Hf8aqDTspFVaC/xt1D1UqlAFgzCezmPD9EoKCgnjlyR5c3bIuAO9N/IXPvluCiFCvRhRvP38HRYsUYv6yjQx7+1vcbqVEsSK8M/QOqleOzNX3dTaBtk1Djm7HixCR5emmx6vq+HTTFYGd6ab/Bi7PVEctABFZBAQDw1T1xyzjO9cCVa2sqlU8/6+cadoniVBEPhKRep73Q3zRRmYul5tBoycxeWx/fpv0HFNmxbJha4I/mg6IeIJEePmuFtz+xi+0Hfw9t7SMplaFsAxlHru5AdOWbqfD0Bk8+O6vvHJXiwzLh9/elF/SJce8cLncvDzuG8aNvI+pHzzJj/NWsmX7ngxlypcNZ8STt9GpfeMz1i9SuBCTxj3OpHGP5ygRulxunnptMpP+8xCLv3yWqWf53j+ftoTwksVZPuUFHurVnuHjvgNgw9YEvpkdy6KJQ5g89iEGjZ6Ey+Vm195DjP9qPnM+GcSiiUNwuZWps2MBGPTqV7w//G7mf/4M3a9vyhv/zfLvNEcCbZsGz+14Xv4H7FPVZule47Op/mxCgJpAO6A38KGIhGe1glcnUESkV2pyEpFKItI0F8FlS1XvT7ff75dkGLtuG9UrRxBdKYLChULo1iGGGfNX+6PpgIinSfUy/LXnCDsSj5LscvPt79u4PqZShjKqULJoIQBKFivE7kMn0pZ1jKnEjsRjbIxPypd41m7aSeUKEVSKKkOhQiFcf1Uj5v2WqadarjS1qkXl60W8ceu3U61SBNEVne/9lg5NmblgTYYyMxesodeNTgeky9WNWbBsE6rKzAVruKVDU4oULkTVChFUqxRB3PrtAKS43Jw8lUxKiosTJ08TFeH5oRHhyLGTABw+epLyERl/gPIi0LbpVEHi3csL8UDldNOVPPPS+xuYpqrJqvoXsAknOZ47vuxaFZF3gPbAnZ5Zx4H3vQo563pLiMgPIrJKRNaKyG0iMk9EmonIK0AxEVkpIhPy2lZWEhKTqFiuVNp0hXKlSEjMnz/s8yGeqFLF2XXg+D/tHzhOVKniGcq8/s1qul9RjbgxtzDhyfY8+/kyAIoXCeHhGy/j9W/z7w9t774kykf+kxjKRYSxd7/3h6dPn07h9gFjufOxd/hl8Vqv10vYeyjj9142nITEQxnLJCZRoazTuQgJCSb0kmIcSDpGQuJZ1t17iAplw3m4zzU0unko9W58jtBLitHes/s8dkhvej3+HvU7P8+kmct49K4OXsea7WcJsG0aAHHGM/Tm5YVlQE0RqSYihYFewLRMZb7F6RUiIhE4u81bs6rUm57hFar6AHASQFUPAIW9iTgbHYFdqtpIVesDafsJqvoMcEJVG6tqn8wrikg/EVkuIssT9yXmQygmK7e0jOarX7cS8/g39HljLu/0uwIRGHRLQ8b/9AfHT6UUdIhpZvxvMF+89SgvP92b1z74np279hdYLIcOH2fGgtXEfTOMdT+M4tiJU0ya6fyQvPflXL4c8xBrp4/k9s6X8/zYbwosTn/Jr0trVDUFeBj4CfgDmKSq60RkhIh08RT7CdgvIuuBucAgVc1yY/Dm0ppkEQnCOWmCiJQB3F6sl501wBsi8iowXVUXervb4zmGMB6gadNmmpcgoiLDiN9zMG16156DREXm3y5LoMeTcPA4FUr/0xOMKl2chIPHM5S5/aoa9H79FwBit+yjSKFgylxShCbVI+jcrArP94whtHhh3KqcSnbx8c+bch1P2YgwdqfrxezZl0TZMlle2JBBOc/uZqWoMjRrWJ0NW+Kp7DnJkZWosuEZv/e9h4iKzHiIKSoyjF2eHmRKiovDR09QOqwEUZFnWbdsOPOXbaRqhTJElCoJQOf2jVi6ZitXt6zDuj930ax+NAC3dIjh1kff8/ozZvtZAmybBueYYX5edK2qM4AZmeYNTfdegSc8L6940zMcB0wBIkVkOPAr8Kq3DZyLqm4CYnCS4igRGZrNKj4RU68qW3Yksj1+H6eTU5g6O45ObRtmv+IFEs/Kv/ZTvVxJqkSUoFBwEF0vj2bWir8zlInff4w29ZwztDWjQilSKJh9R07R9aVZNB/4Lc0HfsuHszbw1vS1eUqEAJfVqsSOXfuI332A5OQUfpq/iqta1vNq3cNHjnP6tNNLPZh0jJXrt1G9Sjmv1m1StwpbdyayfZfzvX8zO5ZObRtkKNOxTQO+/OF3AKb9spI2zWohInRq24BvZsdy6nQy23ftY+vORGLqVaViuVIsX7uN4ydPo6osWLaJWtHlCS9ZnMNHT7B5h3O1x7ylG6kV7V2c3gi0bTpVUJB49Soo2fYMVfVTEYkFrvXMulVVvT8Ycw4iUgE4oKqfi8gh4P5MRZJFpJCqJue1rayEhAQz+qmedB8wDpdL6dOlJXVrRPmyyYCKx+VWhny2jImDriE4SJi4YAsb45N46paGrNx2gFkr/mbYxDhev/dy+l1fF1Xl0Y+W+CyekOBgnnnoZh567iPcLjc3X9ecS6uW591Pf6JerUq0a3kZazfu5ImRn3L46HEW/P4H730+m6kfPMnWnXsZ9fZUgkRwq3Jvz/bUqOpdkgkJCebVgbdy64B3cbmV229qSZ3qUbz8wQ80rluFTm0bcEeXVjw07FOadR9OeGhxPhrlnK2uUz2Km6+N4YpeLxEcHMToQbcSHBxEs/rRdLm6Me3vepWQ4GAa1KrE3V2vICQkmDFDetP3mY8IEiE8tDhvPXfG0aDcf4cBtk1Dzu4uKSji9CazKSTSEGiNs6u8SFXzfMRcRK4HXsPZ5U4GHgJeBwaq6nLP7nMXIO5sxw1TNW3aTBf9vvxciw32dLzs2NPxsnbl5c2IjV2ep4BKV6unHYZ5dy50Ut+YWFVtlpf2ciPbnqGIPAvcDnyDs+v/hYhMUNWX89Kwqv6Ec5AzvXbplj8NPJ2XNowxgSNw0vvZeXMC5S6giaoeBxCRF4EVQJ6SoTHm4nIhDO6akKlciGeeMcZ4xTmbXNBRZC2rgRrG4BwjPACsE5GfPNPX4Vz0aIwx3pHze3DX1DPG64Af0s3/zXfhGGMuVOftbrKq/p8/AzHGXLjO693kVCJSA3gRqAcUTZ2vqrV8GJcx5gIT6D1Db+5A+QT4L05y7wRMAr7yYUzGmAuQePkqKN4kw+KeawJR1S2q+hxOUjTGGK+IQHCQePUqKN5cWnPKM1DDFhF5EGfcsJK+DcsYc6EJ9N1kb5Lh40AJYADOscMw4F5fBmWMufAEeC70aqCG3z1vj/DPAK/GGOM1QQL+uclZXXT9DZ4xDM9GVQPrbntjTOA6D0atyapn+I7fojA+NenZ6ws6hAzu/iSwbmD6cUDrgg7hDEULBxd0CPnuvD1mqKpz/BmIMebCJUDw+ZoMjTEmP533d6AYY0x+uGCSoYgUUdVTvgzGGHNhcob9D+xs6M1zk1uIyBrgT890IxF52+eRGWMuKPn4EHnfxOdFmbeAzsB+AFVdhfNQeWOM8Vp+PTfZV7zZTQ5S1e2ZurguH8VjjLkACRAS4LvJ3iTDnSLSAlARCQYeAfL2cFxjzEUnwHOhV8nwIZxd5SrAHuBnzzxjjPGKyHl8O14qVd0L9PJDLMaYC1iA50KvRrr+kLPco6yq/XwSkTHmgnQhXGf4c7r3RYFbgJ2+CccYcyESKNCBW73hzW5yhiH+ReQz4FefRWSMufAU8DWE3sjN7XjVgHL5HUhB+nnxega/8TUut5s7b76Cx/ted1HFE7dqMx9+9iNut5sO7WLo0SXjKC7fzVjCrLlxBAcHERZagkf+1YWykeEA/G/ibJav/BOAnl3b0qZV/TzH0yK6FA9ffSnBIvywJoEvlp65I9KudiR9r6iKKmxJPMqoHzbQuHI4D7evkVamSunijJi+nl83789xDL/89gdD/zMVl9vN7Te15JE7O2RYfup0CgNGfs7qjTspFVaCD0bcTeWoMmnL/959gKvueJmB93bioduvJn7PQQaM/JzEg0cQhDtubsW/erbLcVzeCrRtGpwxDQOZN8cMD/LPMcMgnIfKP5MfjYvIY8B4VT1+lmV9gWaq+nB+tHUuLpebQaMn8c07D1OhXDhX3/0ando2oE71KF82GzDxuNxuPvhkBsMH30mZ0qEMfP5DWsTUpkqlyLQy1aqW581R/ShSpBAzf17GJxN/5qkBPVi+YhNbtu3mPy89SHJyCs+++D+aNqpJ8eJFch1PkMCj19Zk4OTVJB45xft3xLBoy3627/9nE6kYXow+LSrz8BcrOXoqhfDihQBYufMQ938aC0DJoiFMuK8Fy7YdzPl34nIz5I3JfPWf/kSVDafT/W9wXesG1K5WPq3MxOlLCCtZjCWTnufbn+MY9e73fDCyb9ryYW9/y9Ut66VNhwQH8cIjXWlYuzJHj53k+vtep23zOhnqzC+Btk3D+fGo0CzvQBHnSutGQKTnVUpVq6vqpHxq/zGgeD7VlSux67ZRvXIE0ZUiKFwohG4dYpgxf/VFE8+fW+IpX6405cuWolBIMG1aXsbS2A0ZyjS8rBpFijgJp/alldh/4DAAO+ITuaxOFYKDgyhatDDRlcsSt3pznuKpUz6U+IMnSEg6SYpb+WXDXq6sUSZDmc4No/h25S6OnkoB4NDx5DPquapWJL//dYBTKe4cx7Dij+1EV4qkakXn3+Dma2L4aeGaDGV+XLiWnje0cOJp14iFsZtQdfoMMxespkpUmQyJrlxEGA1rVwbgkhJFqVm1HLsTD+U4Nm8E2jad6ry+HU+df90ZquryvM458nV2RKSEiPwgIqtEZK2IvABUAOaKyFxPmXtEZJOILAWuzG1bOZGQmETFcqXSpiuUK0VCYpI/mg6IePYfOEJEmdC06TKlQ9l/8Mg5y8+et4KmjS4FoFqV8sSt2sKpU8kcPnKcNeu3sW9/3mKNLFmYxCP/jAeSePQUkSUz9jQrlypGpVLFebt3Y969vQktoktlroar60Tyy4a9uYphd2ISFcuGp01HlQ1nd6Z/g92Jh6hQ1mk3JCSY0BJFOZB0jGPHTzHu8zk8eW/Hc9a/M2E/a/78m5jLonMVX3YCbZtOJSJevQqKN8cMV4pIE1Vdkce2OgK7VPVGABEJA+4B2qvqPhGJAoYDTYEkYC5w1jZFpB/QD6BylSp5DMt4a96vq9m8dRcvPd8XgCYNa/Dn1nieHvZ/hIaWoHbNygQFeXO7e94EBwmVShXjsa9WEVmyCG/d1oh7/7eco6ecu0RLlyhM9YgSLM3FLnJevf7xTPrd1o4S5zhUcOz4Ke579mNGDOhGyRJF/RxdwXEeFVrQUWQtq2eghKhqCtAEWCYiW4BjOLv/qqoxOWxrDfCGiLwKTFfVhZl+BS4H5qlqoqf9r4BaZ6tIVccD4wGaNm2W694qQFRkGPF7/vmj2bXnIFGRYXmpMlpvfMsAACAASURBVE/8HU+Z0iXZt/9w2vT+A4cpU+rMJ8GuXLuVyd8t5MXn+lKo0D+bTc+ubenZtS0Ab7wzhQpRZc5YNycSj5zO0BOMvKRIhp6iU+YU63cfxuVWdiedZOfBE1QsVZyNu50ebfvakSz8cx8ud+42jfKRYcTv/WcXNmHvIcpn+jcoHxnOrr0HqVA2nJQUF4ePnaR0WAni1m1n+txVjHx3GoePniBIhCKFQ7i3R1uSU1zc9+zHdLuuGTe2a5Sr2LwRaNt0qvy8A0VEOgJjgWDgI1V95RzlugNfA81VdXmW8WWxbKnn/12A2sANwK1AD8//c0RVNwExOElxlIgMzWkdvhBTrypbdiSyPX4fp5NTmDo7jk5tG1408dSsXpGE3fvZs/cgySkuFv62jhZNa2cos3VbAu/933SefbIX4WEl0ua73G4OH3FObGzbsYdtO/fQpEEN8mLj7sNUKlWM8mFFCQkSrq5TlsVbMp4N/nXzPhpXdnZjw4qFULlUMRIOnUhbfk2dsszZkJjrGBrXqcJffyeyY9d+Tien8N2cOK5vnfEs+fWt6zNphvMnMn3eKlo3rYmI8N17j7Jsygssm/IC/+p5FQPu6sC9Pdqiqjzx8kRqVi3Hg718O+hToG3T8M8JlPw4ZugZI2Ec0AmoB/QWkXpnKVcSeBT4PfOys8lqN1kAVHWLNxVlR0QqAAdU9XMROQTcj/P40ZLAPpyAx4pIGeAwTsJdlR9tZyUkJJjRT/Wk+4BxuFxKny4tqVuj4M66+Tue4OAg+vW9gWGvfo7brVxzVWOqVCrLhK/ncmm1ClzetDb//WI2J06eZvTYyQBERITx3JO9caW4GTzivwAUL1aExx/qRnAe94VcCmPnbOa17g0IChJmrtnNtv3HuefKaDbuPsLiLftZuu0gzaJL88k9zXC7lffnb+XwSedkSvnQIkSWLMKqnbk/ORESEsxLj3en9xPv4XK56dW5JbWrRzH6wxk0qlOZ69s0oHfnljwy8nNa9RxJeGhx3h9+d5Z1Ll29la9/XEbdGlFce/doAAY/cCPXXHFZruPMKv5A2qZT5WPHsAWwWVW3OvXKl8DNwPpM5UYCrwKDvIrvXOdERORv4M1zraiq51x2jvquB14D3EAyzmAPrYCHcY4ltheRe4DBwCFgJXA6u0trmjZtpot+z7L3e9FbsCn3vSRfGDot8zZbsOzpeFm78vJmxMYuz1Mqq1KnoT79f9O8Kvtw62rbcTpIqcZ7Do0BICI9gI6qer9n+k7g8vS5QkRigGdVtbuIzAMGZrebnFXPMBi4BPLnSklV/Qn4KdPs5cDb6cr8F/hvfrRnjAkcQo56hvtUtVmu2xIJwunI9c3JelklwwRVHZHbgIwxJo1ASP5dRBgPVE43XckzL1VJoD4wz3OStjwwTUS6ZNU7zPaYoTHG5FUOe4bZWQbUFJFqOEmwF3B76kJVTQIi0trOh93ka/ISrTHGpJdfl9aoaoqIPIxz2C0Y+FhV14nICGC5qnp3cDKTcyZDVT2Qu1CNMeZM+XlziarOAGZkmnfWy/VUtZ03ddpD5I0xPid49yjOgmTJ0Bjje5K/d6D4giVDY4zPOXegWDI0xpiAvzzFkqExxi8CvGNoydAY4w8FO1ahNywZGmN8zs4mG2OMh51A8TEFUlw5f86Fr4QE4HC+bWtFZl/Ij+YNvKqgQ8igVNvBBR3CGfbNe6mgQ0iTp9GTUwm2m2yMMbabbIwxHtYzNMYY7DpDY4xBgGDrGRpjjF10bYwxgCABvqNsydAY4xfWMzTGXPScS2sCOxtaMjTG+J5Yz9AYYwC7Hc8YYzyDuxZ0FFmzZGiM8Qs7m2yMMdgxwwIxZ8l6nh0zFZfbzR1dWvHoXR0yLD91Opl/D/+cVRt3Ujq0BB+O6kuVCmWY9/sGRr47jeQUF4VCghn2SFfaNKvF8ZOnuW/Ix2yL30dwUBDXta7P0H938Vn8Py9ez+A3vsbldnPnzVfweN/rfNaWxeOda1rU4uUBnQkOCuKzH5bxnwnzMyyvXC6ct5/pTkR4CQ4ePsEDo75iV+JhACqVDWPs092pWDYMVaXnU5+wc/ehHMcwZ8l6hrw5BXfqdn13xs996nQy/Yd/xuoNOykVVoKPRt1DlQplOJB0jHue+T9W/rGdXjdezquDep5Rd5+BH7A9fj+/ThyS47i8Feg9w0AfSCLHXC43z7w+mS/HPMiiiUP4ZlYsG/9KyFBmwrTfCA8tzrKvh/Jg73aMGOc8c7p0eAkmvP4ACyYM5p2hd9B/+Gdp6/y7z9Us+eo5fvn0KZau3srPi9f7LP5BoycxeWx/fpv0HFNmxbJha0L2K/qIxQNBQcJrj3fh1kH/peVdY+h+TSNqVy2bocyI/jfw5U8raH3PW4z+3xyG9uuYtuy9Z3vy9sQFtLxzDNc+8C77Dh7LcQwul5unX5vMV/95iEVfPsvUWbFs3Jp5u15CeMniLJvyAg/2as/wcd8BUKRwCIMfuJFhA245a93T566kRLEiOY4pJ1KPGXrzKigXXDKMW7+d6EqRRFeMoHChELp2iGHmgjUZysxcuIbbbmgBwE3tG7Nw+SZUlYa1K1M+MgyAOtWjOHkqmVOnkyletDCtm9YCoHChEBrWrkzC3pz/snsjdt02qleOILqSE3+3DjHMmL/aJ21ZPN5pWrcyW+P3sz3hIMkpLqbOWcUNretmKFM7uiwL47YAsDBuK508y2tXLUtIcBDzlm8G4NiJ05w4lZzjGOLWb6dapYi07fqWDk3P3K4XrKHXjZcD0OXqxixc5mzXJYoVoWXjGhQtfOaO4NHjp3jvi7k8ec/1OY4pR0QI8vJVUHyeDEUkWkTWppseKCLDRGSeiIwVkZUislZEWniWX+WZt1JEVohIyZy0l5B4iIplw9OmK5QNJyExKUOZ3YlJVCznlAkJCSb0kqIcSMr4a/393JU0rFWJIoULZZifdOQ4s35dS5vmtXISVg7iT6JiuVL/xF+u1Bnx+5PFA1ERocTv/aeNXYmHifL8aKZatzmBzm0vA6Bz28sILVGUUqHFqVE5gqSjJ/l0VB/mf/QIIx7qRFAuuj8Jew9RIf3nLhtOQmLGH+SExKS0bd/ZroudsV1n9vIH0+nf52qKFS2c45hySrx8FZSC7hkWV9XGQH/gY8+8gcC/PfPbACcyryQi/URkuYgs35eYmO9BbdiawMhx03j9mdsyzE9JcdHv+f9xf8+2RFeMyPd2zfnr+XdncGXjasz/6BGubFyN+L1JuNxuQoKDaNUwmufHzeDqB8ZRtUJpbu/UtKDDBWDNpr/ZFr+PG9s18nlbqc9Nvqh7htmYCKCqC4BQEQkHFgFvisgAIFxVUzKvpKrjVbWZqjaLiMw4pH1UZDjx6XZhd+09dMavePnIMOL3OGVSUlwcPnqS0mElPOUPcvfTH/HO0DupVilj3U+88iXVK0fyYK/2ef3c5xQVGUb8noP/xL/n4Bnx+5PFAwn7DlOx7D9tVIgMPXNvY/8R7npuAlfd/zajPpwFwOGjJ9mVmMSazbvYnnAQl8vNjIXraVSrQo5jiCobzq70n3vvIaIiwzOWiQxL2/ad7fpE2nZ9NsvW/MXKP3bQpOsL3NjvP2zZsZcuD43NcWzesp4hpGRqp2i695kfr6Cq+gpwP1AMWCQidXLSWJO6VfhrZyLbd+3ndHIK386Oo2ObBhnKdGxTn69mLAWc3eHWzWoiIiQdOc7tT3zA8/27cHmj6hnWeen96Rw+epIXH++Wk3ByLKZeVbbsSGR7/D5OJ6cwdXYcndo29GmbFk/W4jb8TY1KEVSJKkWhkGC6XdOImYv+yFCmdFjxtJGcH+/TjgkzlqetG3ZJMcp4klKbmOps3LY3xzE0qVuFrTsT2b7L+dzfzI6lY9vM23UDvvzhdwCm/bKSNs1qZTm69L3d27DuhxdZ8e1wfhj/GDWqlGXae4/mODavBXg29MelNXuAsiJSBjgKdAZ+9Cy7DZgrIq2BJFVNEpEaqroGWCMizYE6wAZvGwsJCeblgT3o+ei7uN1uenduSZ3qUbwy/gca16lCx7YN6HNTK/oP/4zmPUZQKrQ440f2BeCjyQv56+99vP7xj7z+sRPi5LH9SU5JYcwns6hZtRxX3/0aAPf1aMOdN1+RD1/PmfGPfqon3QeMw+VS+nRpSd0aUfnejsXjPZfLzVP/mcaU1+8lOEiYMGM5G7btZfC917JyYzwzF/1B68bVGfrA9ajC4lV/MWiMcybX7Vaef3cG3/3nPkSElRvj+d/3y3IcQ0hIMK8MvJVbB7yL263cfpOzXb/8wQ80rluFTm0b0KdLK/oP+5Tm3YcTHlqcD0fdk7Z+k64vcOTYSZKTU5gxfw1fv9Wf2tX9++8Y6LfjiWq+PPsq60acXd5HgXhgK7ANaAesBK4CCgH3qupSEXkbaA+4gXVAX1U9da66Y5o20wWLl/o0/pwIxKfjmazZ0/Gy1qZVc+Jil+cpk9Vt0EQ//W6eV2Vb1AiPVdVmeWkvN/xy0bWqvgW8lX6eiMwDPlfVxzKVfcQfMRlj/CywO4YX5h0oxpjA4hwODOxsWGDJUFXbFVTbxhg/Ow/GM7QDXMYYv8jPk8ki0lFENorIZhF55izLnxCR9SKyWkTmiEjV7Oq0ZGiM8QNBxLtXtjWJBAPjgE5APaC3iNTLVGwF0ExVGwJfA6Ozq9eSoTHGL0S8e3mhBbBZVbeq6mngS+Dm9AVUda6qHvdM/gZUyq5SS4bGGJ/zdhfZkwsjUm+39bz6ZaquIrAz3fTfnnnnch8wM7sY7WyyMcY/vD+Bsi+/rjMUkTuAZjjXM2fJkqExxi/y8dKaeKByuulKnnkZ2xO5FngWuCqrGzdS2W6yMcYv8vGY4TKgpohUE5HCQC9gWsa2pAnwAdBFVb26Gdx6hsYY38vH6wxVNUVEHgZ+AoKBj1V1nYiMAJar6jTgNeASYLLnDPUOVc3yWR2WDI0xfpGfd6Co6gxgRqZ5Q9O9vzandVoyNMb4nBD4d6BYMjTG+EWA58ILIxkGF+Qjtcx5L3HuiwUdwhkiWg8s6BDSnNrwd/5UFOB/phdEMjTGBL5AH9zVkqExxi8COxVaMjTG+EuAZ0NLhsYYn7PBXY0xBs6LwV0tGRpj/CLAc6ElQ2OMP3g3cGtBsmRojPGLAM+FlgyNMb6Xk+ebFBRLhsYY/wjwbGjJ0BjjF3ZpjTHGYMcMjTEGBAJ9PJULJhn+vGQ9Q96Ygsvt5s6bW/HY3ddlWH7qdDIPDfuMVRt2UiqsBB+/eA9VKpQBYMwns/h82hKCg4J4+ckeXNOqLgCNbn6BS4oXITgoiJDgIH759CkA1mz6myde+YpTp5IJCQ7itad70vSy6Pz7LIvXM/iNrz2f5Qoe73td9iv50MUYz5wl63l2zFRcbjd3dGnFo3d1yLD81Olk/j38c1Zt3Enp0BJ8OKovVSqU4UDSMe4d/H+s+GMHvW68nFcH3grA0WMn6fzg2LT1E/YeokfHZrz4ePdcxXfN5bV5+bGuBAcH8dn3v/Ofz37JsLxy+VK8PeQ2IsJLcPDwcR4Y/gW7EpMAGN6/Mx2uqEtQkDBv2SaeGfNtrmLIucDOhgHxDBQR2SYiESISLSJrc7q+y+XmqdGTmTT2IZZ89SxTfoplw9aEDGU+n7aE8JLFiZ36Ag/1bs+wd74DYMPWBKbOimXxl0OYPPYhBo2ehMvlTltv2nsDWDDhmbRECPDC29/x1P0dWTDhGQY/cCPD3v4utx/9rJ9l0OhJTB7bn98mPceUWWd+Fn+6GONxudw88/pkvhzzIIsmDuGbWbFs/CtjGxOm/UZ4aHGWfT2UB3u3Y8Q45xEcRQqH8Ey/Gxn+SNcM5S8pUZR5nz2d9qpUvhQ3tmuUq/iCgoTXBnbj1ic/pOXto+l+bRNqR5fLUGbEwzfx5czltL7rDUb/dzZDH7oBgBb1o7m8YTSt73qdK+54jSZ1K3Nlkxq5iiMnUgd3zadnoPhEQCTDvIpdt51qlSKIrhhB4UIhdLuuKTMXrMlQZsb8NfS68XIAbr66MQuWbUJVmblgDd2ua0qRwoWoWjGCapUiiF23Pcv2BDhy7CQAh4+eoHxEWD5+lm1UrxxBdCXPZ+kQw4z5q/Otfosne3HrtxNdKTJte+raIeaM7WnmwjXcdkMLAG5q35iFy53tqUSxIrRsXIMihQuds/4tO/ay7+BRWjXOXRJqWq8KW//ez/ZdB0hOcTH15xXc0OayDGVqR5djYexmABbGbqZTm/oAKEqRwiEUDgmmSKEQCgUHk3jgSK7iyKkcPDe5QPglGYrIIBEZ4Hk/RkR+8by/WkQmnGOd6iKyQkSaZ1d/QuIhKpYrlTZdoWw4CYmHMpVJomK5cABCQoIJvaQYB5KOZbmuAN0fGUf7u0bzyTeL0sq89ER3XnjrO+p3fp6hb33L0H9n+ZyZHHHiTBdPuVIkeHZvCsLFGE9C4iEqlg3/p42y4We0sfuM7akoB5KOeVX/N7Nj6XptTK7vyIiKDCN+zz/b967EJKIiM/4gr9u8i87tGgDQ+aoGhJYoSqnQ4ixbu52FcVvY8P0wNnz/Ar8s3cim7V49PC7PrGfoWAi08bxvBlwiIoU88xZkLiwitYEpQF9VXXaW5f1EZLmILN+3L9FnQc/48HHmffY0k/7zEP83eQGL45xf2v9O+ZUXH+/G2ukjGfVYNwaMOms+N+asvpkdR7frYnzaxvPvfM+Vjasz/5MnuLJJdeL3HsLldlOtYhlqR5flsq4jqHfzCNo0vZRWjar5NJZUIuLVq6D4KxnGAk1FJBQ4BSzBSYptcBJlepHAd0AfVV11tspUdbyqNlPVZhERkURFhhO/52Da8l17DxEVGZ5hnfS/pikpLg4fPUHpsBJZrlvB0zuILF2SG9s1Ina9s/s88Yffuam9c7yn67VNiF2/Ixdfydk5caaLZ8/BM371/elijCcqMpz4vel6XnsPndFG+TO2p5OUDiuRbd1r/4wnxeWmUZ0quY4v/V4OQIXIsDN7rvsOc9eQ/3FV3zcZ9cFMAA4fPUnnqxqwbO12jp04zbETp/l5yQaa14/OdSw5YbvJgKomA38BfYHFOAmwPXAp8Eem4knADqC1t/XH1KvC1p2JbI/fx+nkFKbOiqVjmwYZynRq24Avf/gdgO9+WUmbZrUQETq2acDUWbGcOp3M9vh9bN2ZSNPLqnLsxKm044LHTpxi7u8bqFsjCnD+EBZ5eokLlm2iRuXInH0hWX6WqmzZke6zzI6jU9uG+Va/xZO9JnWr8NfORLbv2s/p5BS+nR13xvbUsU19vpqxFIDv566kdbOaXvVqps6Kpdt1TfMUX9wfO6lRKYIqUaUpFBJMt2ubMPPXdRnKlA4rkRbP43ddw4TpTqx/7znElU1qEBzsXCFxZZMabNq2J0/xeMPbXeSC3E3256U1C4GBwL3AGuBNIFZVNdNGdBq4BfhJRI6q6hfZVRwSEszoQbfSY8C7uNxKn5taUrdGFC998ANN6lahU9sG3NGlFQ++8ClNuw2nVGhxPnrxHgDq1oii67UxtLrtJUKCgxj91K0EBweReOAIdw76EIAUl5se1zfj2lb1ABg7pDeD35xCSoqLIkUKMWZwr3z7kkJCghn9VE+6DxiHy6X06dIyLQkXhIsxnpCQYF4e2IOej76L2+2md+eW1KkexSvjf6BxnSp0bNuAPje1ov/wz2jeYwSlQoszfmTftPVjug7jyPGTnE5OYeb81Ux+qz+1qzkxTpuzgolvPpin+FwuN0+9OZUpY/oRHCxMmL6UDX/tYfD917Nyw9/M/HUdrWNqMPTBG1CFxSu3MuiNKQB8N3cVbZteyqLPBqKqzPl9Iz8uWp+neLwV6HegiKr6pyGRa4AfgXBVPSYim4D3VfVNEdmG51giMF1V64tIODAbGKmq085Vb0zTZrrotzMOKxaYQB+myJwpJd2lVIEiss2ggg4hzal1E3Af252nDbtxTFOdveB3r8qWLVkoVlWb5aW93PBbz1BV5wCF0k3XSvc+2vN2H1DfM+8QkO2ZZGPM+SHQuwkXzB0oxphAJvaoUGOMSb0DJZBdEHegGGNMXlnP0BjjF4HeM7RkaIzxi0C/tMaSoTHG9+y5ycYYc36cQLFkaIzxC9tNNsYYAr9naJfWGGP8Ij9HrRGRjiKyUUQ2i8gzZ1leRES+8iz/XUSis6vTkqExxj/yKRuKSDAwDugE1AN6i0i9TMXuAw6q6qXAGODV7Oq1ZGiM8TkBgkS8enmhBbBZVbeq6mngS+DmTGVuBv7nef81cI1kM4rKeX/McEVc7L7ihYOyfmiJdyJwBooIFIEWDwReTBZP1vIrnqp5rSAuLvanYoUkwsviRUVkebrp8ao6Pt10RWBnuum/gcsz1ZFWRlVTRCQJKEMW38d5nwxVNV9GVhWR5QUxbNC5BFo8EHgxWTxZC6R4VLVjQceQHdtNNsacb+KByummK3nmnbWMiIQAYcD+rCq1ZGiMOd8sA2qKSDURKQz0AjIPAD0NuNvzvgfwi2YzkvV5v5ucj8ZnX8SvAi0eCLyYLJ6sBVo8+cJzDPBh4CcgGPhYVdeJyAhguWdk/P8DPhORzcABnISZJb8N+2+MMYHMdpONMQZLhsYYA1ykyVBEhonIwIKOI5WILC7oGLIjIh+lXuUvIkMKOJbHRKT4OZb1FZF3/B1TIBKRbSISISLRIrK2oOMJdBdlMgw0qnpFQceQHVW9X1VTH7BboMkQeAw4azI0JrcummQoIs+KyCYR+RWoDQSJSKxnWSMRURGp4pneIiLFReRWEVkrIqtEZIEPYzsqjtc87a0Rkds8y4JE5F0R2SAis0Vkhoj08FUsnjZLiMgPns+9VkRuE5F5ItJMRF4BionIShGZ4Ms4zhHLC0AFYK6IzPWUucfzb7sUuNKHsWToYYnIQM9exjwRGev5TtaKSAvP8qs881aKyAoRKZnP8QwSkQGe92NE5BfP+6vP9W8jItU9sdhjeDO5KC6tEZGmOKfWG+N85jggFue2n1CgDbAcaONJlntV9biIDAWuV9V4z0PtfambJ75GOLdRLfMk4CuBaJwb0ssCfwAf+ziWjsAuVb0RQETCgIcAVPUZEXlYVRv7OIasYrkHaK+q+0QkChgONAWSgLnACj/Fll5xVW0sIm1x/n3qAwOBf6vqIhG5BDiZz20uBJ4E3gKaAUVEpBDO9py67aQRkdo49/H2VdVV+RzLee9i6Rm2Ab5R1eOqeph/LtBcjLPBtAVe8vy/Dc5GBrAI+ERE/oVzPZMvtQYmqqpLVfcA84HmnvmTVdWtqrtx/th9bQ3QQUReFZE2qprkhzZzG8vlwDxVTfTctP+V/0MEYCKAqi4AQj0/nouANz29t3BVTcnnNmOBpp4f9FPAEpykmH4bThUJfAf0sUR4dhdLMjyXBTgbTlWcDaURTvJZCKCqDwLP4dzWEysiZQooTr9S1U1ADE4iGuXpIV/0sQApZPybKZrufeYLdlVVXwHuB4oBi0SkTn4Go6rJwF9AX5wf9oVAe+BSnD2I9JKAHTjbtzmLiyUZLgC6ikgxz3GbmzzzFwJ3AH+qqhvnSvUbgF8BRKSGqv6uqkOBRDLeD5nfFgK3iUiwiETi9FKX4vQuunuOHZYD2vkwBgBEpAJwXFU/B17DSUbpJXt2x3zuHLEcAVKPv/0OXCUiZTwx3erDcPYAZT1tFQE6p1uWeoy3NZCkqkme7WeNqr6KcwtZviZDj4U4u+MLPO8fBFac5daz08AtwF0icrsP4jjvXRTHDFU1TkS+AlYBe3E2TFR1m4gIzoYEThKspKoHPdOviUhNnOHY5njW90mIwDdAK08bCjylqrtFZApwDbAeZ0iiOJxfeV9qgPPZ3UAyzvHC19MtHw+sFpE4Ve1TALG0An4UkV2q2l5EhuHsIh4CVvoqEFVNFueWr6U4AwFsSLf4pIisAAoB93rmPSYi7QE3sA6Y6YOwFgLPAktU9ZiInOTMXeTU+I+JSGdgtogc9dy2ZjzsdrwC5tn1jlPVc44ZJyKXqOpRT9mlwJWe44cmAIjIPGCgqi7PrqwJXBdFzzBQeXYB55Gx13U20z0H5AsDIy0RGpP/rGdojDFcPCdQjDEmS5YMjTEGS4bGGANYMrzgiYgr3T2zk+Uco714WVc7EZnued9FzvLw7nRlw0Wkfy7aOOuIQuean6nMJzm5bzvzvcbm4mbJ8MJ3QlUbq2p9nAtvH0y/UBw53g5UdZrnDotzCQdynAyNKSiWDC8uC4FLPT2ijSLyKbAWqCwi14nIEhGJ8/QgLwEQkY7ijJgThzOYBJ75aeMGikg5EflGnJFlVonIFcArQA1Pr/Q1T7lBIrJMRFaLyPB0dWUeUShLIvIvTz2rRGRKpt7utSKy3FNfZ0/5YHFGBEpt+4G8fpHmwmPJ8CIhzuMSO+Hc4wtQE3hXVS8DjuHcg32tqsbgjODzhIgUBT7EuX2xKVD+HNW/BcxX1UY4t8utA54Btnh6pYNE5DpPmy1wRudpKiJtJeOIQjfgDE6Rnamq2tzT3h/AfemWRXvauBF43/MZ7sO5Ra65p/5/iUg1L9oxFxG76PrCV0xEUm9RW4jz1LAKwHZV/c0zvyXOEGGLnLsTKYxze1sd4C9V/RNARD4H+p2ljauBuwBU1QUkiUipTGWu87xSh9e6BCc5lsQzopCnDW9uEasvIqNwdsUvwXlKWqpJnvvM/xSRrZ7PcB3QMN3xxDBP25u8aMtcJCwZXvhOZB570JPwjqWfBcxW1d6ZyuXnmIUCvKyqH2Rq47Fc1PUJ0FVVV4lIXzIOXnHG6DGeth9R1fRJExGJzkXb5gJlu8kG4DfgShG5FNJGl66FMxBBtIjU8JTrfY715/D/7dyhSkRhEIbh98uiNotNQfBuLDaTgrjJG9AbEbtXIF6BbLOv0WLYKmKRl2FmDQAAAMxJREFUMfwTVnF1+75POeXAHE4Y5h/O+Tr8tfdz23xPloExvZ0u7CJ3k+ywPFHoL5vAa6fU/AyKOM5I+NkH9oBZ1570/SQ5SLKxQh2tESdDUVXznrDuMqKpAK6q6jnJOXCf5J1xzP4tuv4SuElyBnwCk6qaJnnsT1ceem94CEx7Mn0DTpYlCv3jmhHdNe/r4jO9MMIstoCLqvpIcsvYJT5lFJ8DR6u9Ha0L/02WJDwmSxJgM5QkwGYoSYDNUJIAm6EkATZDSQJshpIEwBeSXDine2EnTAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_confusion_matrix(clf, X_test, y_test,\n",
    "                      display_labels=label_encoder.classes_,\n",
    "                      cmap=plt.cm.Blues,\n",
    "                      normalize=\"true\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Classification with SVM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "import sklearn.svm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/sergio/PycharmProjects/MLEARN/venv/lib/python3.7/site-packages/sklearn/svm/_base.py:947: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Pipeline(memory=None,\n",
       "         steps=[('columntransformer',\n",
       "                 ColumnTransformer(n_jobs=None, remainder='drop',\n",
       "                                   sparse_threshold=0.3,\n",
       "                                   transformer_weights=None,\n",
       "                                   transformers=[('acceleration',\n",
       "                                                  Pipeline(memory=None,\n",
       "                                                           steps=[('fouriertransform',\n",
       "                                                                   FourierTransform(window=None)),\n",
       "                                                                  ('pca',\n",
       "                                                                   PCA(copy=True,\n",
       "                                                                       iterated_power='auto',\n",
       "                                                                       n_components=32,\n",
       "                                                                       random_state=None,\n",
       "                                                                       svd_solver='auto',\n",
       "                                                                       tol=...\n",
       "                                                                       svd_solver='auto',\n",
       "                                                                       tol=0.0,\n",
       "                                                                       whiten=False))],\n",
       "                                                           verbose=False),\n",
       "                                                  slice(256, 512, None)),\n",
       "                                                 ('metadata', 'passthrough',\n",
       "                                                  slice(-4, 516, None))],\n",
       "                                   verbose=False)),\n",
       "                ('linearsvc',\n",
       "                 LinearSVC(C=1.0, class_weight=None, dual=True,\n",
       "                           fit_intercept=True, intercept_scaling=1,\n",
       "                           loss='squared_hinge', max_iter=1000,\n",
       "                           multi_class='ovr', penalty='l2', random_state=None,\n",
       "                           tol=0.0001, verbose=0))],\n",
       "         verbose=False)"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf = sklearn.pipeline.make_pipeline(\n",
    "    features,\n",
    "    sklearn.svm.LinearSVC()\n",
    ")\n",
    "clf.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "         dws       0.64      0.92      0.75       236\n",
      "         jog       1.00      0.34      0.51       265\n",
      "         sit       0.53      0.97      0.69       663\n",
      "         std       0.82      0.46      0.59       615\n",
      "         ups       0.86      0.66      0.75       298\n",
      "         wlk       0.98      0.77      0.86       666\n",
      "\n",
      "    accuracy                           0.71      2743\n",
      "   macro avg       0.80      0.69      0.69      2743\n",
      "weighted avg       0.79      0.71      0.70      2743\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_train, clf.predict(X_train), target_names=label_encoder.classes_))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<sklearn.metrics._plot.confusion_matrix.ConfusionMatrixDisplay at 0x12bd8f610>"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAUMAAAEGCAYAAAD/pvfYAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAAgAElEQVR4nOzdd3gU1RrA4d+X3ZCebHohoUqVmiAdpEiTpqCIoIK9IRYEy1VREb2AWMEu1wIWUJQu0ouAQkIv0iSBJIQESCBASLJ77h8blmwCYdMXPC/PPE9m5sycb2eWb8+0M6KUQtM07d/OpbID0DRNcwY6GWqapqGToaZpGqCToaZpGqCToaZpGgDGyg6gtMTVS4m7qbLDsGleN7yyQygk1+JcdwwYXaSyQ9CKIT7+EGlpaaXaaQbf6krlnnOorDqXulgp1bM09ZXE1Z8M3U24xTxS2WHY/LH05coOoZATmdmVHYKdAO8qlR2CVgztWrUo9TpU7jnc6g1yqGzWlqlBpa6wBK76ZKhp2tVAQJz7rJxOhpqmlT8BXAyVHUWRdDLUNK1iiHOfK9bJUNO0CqAPkzVN06x0y1DTtH89QbcMNU3TrIfJumWoaZqmryZrmqbpCyiapmmQd85QHyZrmqbplqGmaZo+THYSXW+ozVuP9cDgIny7aDPv/bDObn5UiB8fPtuXIJMnJ0+f4+G3fiUp7TSNaocy+cmb8fF0w2KxMPm7tfyycle5x7t03S5emPwTZouFu/u35enh3cu8jlV/7WbclF8xmy3c0bs1jwzpajf/fHYuz771HTv2Hsbf14sPxt5DZFgAOblmXpj0Izv3HcFstnBr9xY8OvQmADoOHoeXpxsGFxcMBhfmfPpMmccNFbN9rvaYnC0eBDDoCygAiMirQKZS6u2KqhPAxUWY9ERPbn1uBkmpp1g+9QEWrdvL3wlptjKvP3wTPyzZxg9LttGhWQ1eub8Lj0yYw7msHB6dMIeDiScIC/RmxUcPsGzjAU6dOV9u8ZrNFkZPnMkvU0YQEWqiy7BJ9OrYmPq1yq5rMLPZwqvvz+brSY8QFuzHrY+8S9e211OnRpitzKyFf+Ln48GKGf9h3vLNTPh0Ph+OvYdFK7eQnZPLomljOJeVTY/hE+jbNZrIsAAAZrz7GAF+3mUW66ViL+/tc7XH5Gzx2Dj5OUPnbreWgZh6ERxMOkl8cjo5uRZmr9zJze3q2ZWpVz2YNVsOAbBmyyF6tbXOP5B4goOJJwA4ejyTtPSzBJm8yjXe2J2HqBUVRI3IIKq4GhnQLZqFq7aVaR1b9yRQPSKIahGBVHE10qdLc5b+scOuzNI/djCgxw0A9LqxCevj9qGUAhHOZWWTazaTdT4HV1cj3p5uZRpfUSpi+1ztMTlbPFZ5h8mODJWkXGsWkf+IyF4RWQvUA1xEJDZvXlMRUSJSLW/8gIh4isjtIrJDRLaKyOrSxhAe5EvisVO28aTUU4QH+tiV2XkwhT7t6wPQp319fL3c8Pf1sCsTXS8CV6OBf5JOlDakIiWnZlA11N82HhHqT3JqRpnWkZKWQXjIxQ5xw4JNpKTZ13E0XxmjwYCPtzsnT52h141N8XCvQpuBr9Jh8DgeGNQJk6/1B0JEGD76U/o99A7fz1tfpjFfUBHbp7icLSZni8dGxLGhkpTbYbKIxACDgWZ59cQBsYC7iPgCHYBNQIe8ZHlMKXVWRF4BeiilEkXkkl1Yi8hDwEMAuPmVOtaXP13CxBG9GNKjKeu2xZOYegqz2WKbHxrgzSfP38JjE+fwb3/N9NbdCRhcXFj306tknD7L4Cen0C6mLtUiAvnxgxGEBZtIO3maYc9+Qu1qIbRsWruyQ9acxb/4AkoH4Bel1FkAEZmbN30d0A7oCLwJ9MR6enVN3vw/gK9EZCYw+1IrVkp9BnwG4OJTtcj0lJx2iqohvrbxiGBfko+ftitz9Hgm97w2CwAvd1f6dmhgOy/o41mFH8cP5o1pK9i0O9GxT14K4cF+JKactI0npZwkPLj0CT+/0CA/ko+l28aPpqYTGmRfR1hemfBgE7lmM6czs/D39WLessV0bFkfV6OBIH8fYq6vyfa/D1MtIpCwYOtvV5C/D907NGbrnoQyT4YVsX2Ky9licrZ4gEpv9TmiMlL1aqyJsjowB2gKtCcvGSqlHgFeAqKAWBEJLE1lcX8nUbtqANXCTLgaXRjQ6XoWrdtrVybA18O2n56+sz0zftsCgKvRhW9fHcQPS7Yxd83u0oThsOiG1TmQkEp8YhrZObnMXhJHr45NyrSOJvWjOJSYyuHk42Tn5DJ/+Wa6tm1kV6Zr2+uZvXgjAItWbaNN8+sQESJCTazfvA+As+fOs2V3PLWrhXD23Hkyz2bZpq/ZtJe6NcMoaxWxfa72mJwtHhsXg2NDJSnPluFqrC28t/Lq6Qt8ijXpjQdWK6UsInICuBl4AUBEaiul/gT+FJFeWJPi8ZIGYbYoxnz4Gz//dwgGF2HGb1vZE5/KC8NuZMveZBat30v7pjV45f7OKGDdtgRGf7gIgFtvvJ62TaoR4OvBkO5NAXhs0lx2HEgpaThXZDQamDhmEANHTsVsVgzt15oGtcv2KqDRYGDsyAEMH/MZFouF23q1pG7NMN6dtojG9aK4qV0jBvVuxag3v6Pz0PGYfD15/+V7ALjrlvY8N+EHeg6fgAIG9ryB+rUjSEg6zqMvTwOsVzP73hTNjS0blGncUDHb52qPydnisXL++wxFleNJMBH5DzAMOAYkAHFKqbdF5DAwTin1mYi8CAxWSjXJW2Y2UAfrofMy4ClVRJAuPlWVM70Q6qR+IdQV6RdCXV3atWpBbOymUh3juvhVU27tn3WobNbCJ2OVUqV/C1Uxlet9hkqp8VhbgQWnR+X7+02s5w4vjA8oz5g0TasEuj9DTdM0uBoOk3Uy1DStYuj+DDVN03D6W2t0MtQ0rfyJPkzWNE2z0i1DTdM067PrzkwnQ03Typ2113+dDDVN+7cTQVx0MtQ0TdMtQ03TNNDJUNM0DdDJUNM0Le8KSmUHUbSrPhk2qh3G3J9GV3YYNrO3HansEArZfvRMZYdg54m2NSs7BDu6F53yJ0iZtgxFpCfwPmAAvlBK/bfA/GrA14Apr8zzSqmFRa3zqk+GmqZdHVxcyuYJFBExAFOBbsARYKOIzFVK5X+P70vATKXUxyLSEFgI1CgyvjKJTtM07QpExKHBAS2B/Uqpg0qpbOAHoH+BMgq48L4PPyDpSivVLUNN08pf8c4ZBonIpnzjn+W99+iCqsDhfONHgFYF1vEq8LuIPAF4ATddqVKdDDVNqxDFOGeYVgY9Xd8JfKWUmiwibYBvRaSRUspyuQV0MtQ0rdyV8QWURKzvRrogMm9afvdjffMmSqn1IuIOBGF9Bckl6XOGmqZVCHERhwYHbATqiEhNEamC9f3scwuUSQC6AohIA8AdSC1qpbplqGla+ZOyu+laKZUrIiOAxVhvm5mmlNopIq8Dm5RSc4FRwOci8jTWiynDi3qxHOhkqGlaBSnL+wzz7hlcWGDaK/n+3gW0K846dTLUNK1C6MfxNE371yvrJ1DKg06GmqZVDOfOhToZappWAaTsHscrLzoZappWIfRhsqZpGujDZGewZuMe3vxoDhaLhdt6teLBwV3s5m/cdoC3Pp7L3oPJTP7PUHp0bArA7v2JvPbBbDLPZmFwceHhIV25uVOzUsezfftBvv9+KUpZ6NChKTff3MZu/sqVm1m+PA4XF8HNrQrDhvUkIiLINv/48QxefvkL+vVrT8+eBR/JLL5De+NZtWA1FouiUYuG3HDjpZ+E2rdjPwu+X8Sdjw4iNDKUjJOn+Oa96fgH+QMQHhVG11s6lyiGVX/tZtyUXzGbLdzRuzWPDOlqN/98di7PvvUdO/Yext/Xiw/G3kNkWAA5uWZemPQjO/cdwWy2cGv3Fjw69CYOJhxj5Ovf2JY/nHycp+7tyb233Vii+K5k6bpdvDD5J8wWC3f3b8vTw7uXSz1XazygW4aFiMg6pVTbiqrPbLYw7sNf+HLCQ4QG+TFoxPt0btOQ66qH2cpEhPjz1ug7mDZrld2y7u5V+O+YwdSIDOZYWgYDH3+P9i3q4evtUeJ4LBYLM2b8zqhRg/H392HcuK9o1qyOXbJr1aohnTo1B2DLln38+OMynn76Dtv8H39cTqNGtUocQ8F4VsxbyYB7b8Hb15vvP/6RWg1qERgSYFcu+3w2W9ZvJSwq1G66KcCPu564s1QxmM0WXn1/Nl9PeoSwYD9ufeRdura9njo1Lu6jWQv/xM/HgxUz/sO85ZuZ8Ol8Phx7D4tWbiE7J5dF08ZwLiubHsMn0LdrNLWqhTD/i2dt6297+2t0b9+4VHEWFf/oiTP5ZcoIIkJNdBk2iV4dG1O/Vni51He1xQMUp0eaSlPhZzQrMhECbPs7gWoRgUSFB1LF1cjNnZqxfN1OuzJVwwKoVysClwI7q2ZkMDUigwEICfIj0OTNifTMUsVz8GAyISH+BAebMBoNtGzZkM2b99mV8fBws/19/nwO+Y8v4uL2EhTkR9WqQZSFo0dS8Asw4Rfgh8FooG6TuhzYfbBQuXVLN9CiQzQGY9n/fm7dk0D1iCCqRVj3UZ8uzVn6xw67Mkv/2MGAHjcA0OvGJqyP24dSCkQ4l5VNrtlM1vkcXF2NeHu62S27Lm4f1SICqRpmn+DLSuzOQ9SKCqJGZBBVXI0M6BbNwlXbyqWuqzGeC8qwC69yUeHJUEQyxWqSiOwQke0ickfePBcR+UhE9ojIEhFZKCK3laa+Y2kZhAWbbOOhQSZS0jKKvZ5texLIyTFTLSKwNOGQnn6agAAf27i/vw/p6acLlVu+PJbnn/+EWbNWMGSItfehrKxsFi3aQL9+7UsVQ35nTp3Bx8/bNu7j682ZDPuEfyzxGJkZmdSsX7iH6oyTp5gx5Xtmff4ziYcKPivvmJS0DMJDLu6jsODC++hovjJGgwEfb3dOnjpDrxub4uFehTYDX6XD4HE8MKgTJl8vu2XnL99M367NSxSbI5JTM6ga6m8bjwj1Jzm1+N+xazWeC8rw2eRyUVnnDAcAzYCmWHuS2Cgiq7E+PlMDaAiEALuBaQUXFpGHgIcAIiKjCs4uc8eOn+K5Cd/z1ujBFXZ7QJcuMXTpEsOGDTuZP38d99/fhzlz1tK9+w24u1dcN/XKoli1aC3dBxbuDs7Lx4v7xwzHw9ODlMRjzJuxgLtHDsWtAuPbujsBg4sL6356lYzTZxn85BTaxdS1/Whl5+SybN1ORj/Yu8Ji0i7N2Q+TKysZtge+V0qZgRQRWQXckDd9Vl6fY0dFZMWlFs7r6PEzgCbNYop8+DokyI+jqem28ZS0dEKD/BwONPNMFo+89CVP3duTZg2rO7zc5ZhMPpw4cbElePLkaUwmn8uWb9myIdOn/w7AP/8kERu7h1mzVnD27HlEBFdXI127xpQ4Hi9fL07nawmePpWJV76WYnZ2NsdTjvPTF7MBOJt5lrnTF9Dvrt6ERoZiNFrPn4ZWDcEvwI/0tJOERtqfV7yS0CA/ko9d3EdHUwvvo7C8MuHBJnLNZk5nZuHv68W8ZYvp2LI+rkYDQf4+xFxfk+1/H7Ylw1V/7uH6ulUJCrj8Ni6t8GA/ElNO2saTUk4SHuz4d+xajwco044ayotz3wVZBhrXiyI+MY0jycfJzsll4cotdG5zvUPLZufk8sSrX9G/W4ztCnNp1awZTkrKCVJT08nNNfPXX7to1uw6uzIpKSdsf2/btp+QEOshz/PP38XEiY8xceJjdOvWgt6925QqEQKEVQ0l/Xg6GScyMOea2bttL7XzHQ67ubvxyH8e5P7Rw7l/9HDCosJsifDsmXNYLNa+MjNOZJCelo5fQPH/0zWpH8WhxFQO5+2j+cs307VtI7syXdtez+zFGwFYtGobbZpfh4gQEWpifd4517PnzrNldzy1q4XYlpu3PI6+XaKLHVNxRDeszoGEVOIT08jOyWX2kjh6dWxSrnVeTfFAXkfX4thQWSqrZbgGeFhEvgYCgI7AaMANGJY3PRjoBHxXmoqMBgMvjbiVB174HItFMaDHDdSpEcYHX/1Go7pRdGl7Pdv/TuCJV7/mVOZZVmzYxYff/M78L0bz26qtbNp+kPRTZ/l1sbUX8jdH30GD66qWOB6DwYWhQ7vz7rs/YrEo2rdvQtWqwfz662pq1AinWbM6LFsWy+7d8RgMLnh6unP//eV3iOdicKFz3xv55au5KGXh+uiGBIYGsn7pBkKqhlC7weWvWif+k8j6ZX/i4uKCiNC1f2fcPd2LHYPRYGDsyAEMH/NZ3u1PLalbM4x3py2icb0obmrXiEG9WzHqze/oPHQ8Jl9P3n/5HgDuuqU9z034gZ7DJ6CAgT1voH7tCMCaHP+I3cv4Z24v0bZxOH6jgYljBjFw5FTMZsXQfq1pULvyrtw6WzxWzn81Wa7QxVfZVyhyGuuLWiYCvbD2NfaGUupHEXEBPsKaBA9j/UGZoJRacrn1NWkWo+Yu/aPc43bUuoS0yg6hEP2q0KLpV4UWrV2rFsTGbipVJnMPq6uqD/vQobJ7J/aMLYNu/4utQluGIhIInMjrZHF03mCjlLKIyLNKqcy8sn8B2ysyRk3TykElHwI7osKSoYhEACuBt69QdL6ImIAqwDil1NHyjk3TtPIlgEsl3jbjiApLhkqpJKCuA+U6lX80mqZVNN0y1DRNw/lvrdHJUNO08qfPGWqaplm7/dedu2qapqFbhpqmaYA+Z6hpmqbPGWqapsGFZ5OdOxvqZKhpWoVw8lyok6GmaRVDP4GiaZp2FfRneNUnQ1eDEGYqfrdR5WWAKbKyQyjk/vtHVHYIdl5Y/0Flh6BVsAv9GTqzqz4Zapp2NXD+/gx1MtQ0rUI4eS7UyVDTtAog+gKKpmmavs9Q0zTtAp0MNU3T0OcMNU3TAOdvGTp3B2Oapl0bHHxnsqP5UkR6isjfIrJfRJ6/TJlBIrJLRHaKyBVfOaxbhpqmlTtr565l0zIUEQMwFegGHAE2ishcpdSufGXqAC8A7ZRSJ0Uk5Err1S1DTdMqhIuIQ4MDWgL7lVIHlVLZwA9A/wJlHgSmKqVOAiiljl0xvmJ+Hk3TtBIpxmFykIhsyjc8VGBVVYHD+caP5E3Lry5QV0T+EJENItLzSvHpw2RN08qdFK+jhjSlVItSVmkE6gCdgEhgtYg0VkqlF7XAJYmIb1E1KaVOlTBITdP+hcrwAZREICrfeGTetPyOAH8qpXKAf0RkL9bkuPFyKy2qZbgTUFhvHr/gwrgCqjkcupNbum4XL0z+CbPFwt392/L08O7/qni6tmnAW6Nuw+Diwrdz1vHe10vs5keF+fPhK3cRZPLm5KmzPPzK1yQdS6d9TB3efGagrVyd6qHc/5//sXDVtmLHsHz9Ll56bzZms4Wh/dow8p5udvPPZ+cw4vXpbNtzGH8/Lz57YzjVwgNZ9dce3vhoLtk5Zqq4GnhlxC10aFGXzDNZ9Hv0fdvyycfSGdijBW88PbBg1WXi3/4dckQZPo63EagjIjWxJsHBwJACZX4F7gT+JyJBWA+bDxa10ssmQ6VU1OXmlRcR+QJ4Rym1S0ReVEq9Wd51ms0WRk+cyS9TRhARaqLLsEn06tiY+rXCy7tqp4jHxUWYNGYQt46YQlJKOsu/Hs2i1dv5+5+jtjKvP3krPyz4ix8W/EmHFnV55fF+PDL2G9bG7qPj0P8CYPL1JG72WFZs2F3sGMxmC89PnsXM9x8nIsREj/vepkeHRtSrefEzfzdvAyYfT/786RV+WRLLuKlz+fyNewnw8+LbSQ8TFuzH7gNJDH7qY7bOG4e3lzvLv3nOtny34RPp3alpKbZU0fH/m79DjhCsV5TLglIqV0RGAIsBAzBNKbVTRF4HNiml5ubN6y4iuwAzMFopdbyo9Tp0AUVEBovIi3l/R4pITGk+zOUopR7Id3n8xfKoo6DYnYeoFRVEjcggqrgaGdAtukQtm6s1npjra3DwcBrxicfJyTUze0kcN9/YxK5MvVrhrNn0NwBrNu2lV8fGhdbTv2tzlq7fxbnzOcWOIW5XPDUjg6lR1fqZb7kpmt9Wb7cr89ua7Qy6uSUAfTs3Y+2mvSilaFwvirBgPwDq1won63wO57PtYziQcIy0k5m0bla72LE54t/+HXKUizg2OEIptVApVVcpVVspNT5v2it5iRBl9YxSqqFSqrFS6ocrxnelAiIyBegM3J036SzwiWMhF7leLxFZICJbRWSHiNwhIitFpIWI/BfwEJEtIjKjtHUVJTk1g6qh/rbxiFB/klMzyrNKp4onPNiPxJSTtvGklJOE5yWXC3buTaRP52YA9OncFF9vD/z9vOzKDOgWzc+LY0sUw9HUdCJCTLbxiBATRwt8Zut2sZYxGg34eLtzIuOMXZn5K7bQuF4kblVc7ab/uiSW/l2jy+0JiH/7d8ghYu3P0JGhsjjSMmyrlHoYyAJQSp0AqpRB3T2BJKVUU6VUI+C3CzOUUs8D55RSzZRSQwsuKCIPXbjsnpqWWgahaEV5+f1faBd9HaumP0e76OtITDmJ2WyxzQ8N9KXhdREsW7+riLWUrz0Hkxn30Vzefu6OQvN+XRrHrd2jKyEqLb+yfAKlPDhya02OiLhgvWiCiAQClqIXcch2YLKITADmK6XWOPqroJT6DPgMICamhSpNEI60jCpSRcfjSCviaFoG94z5AgAvjyr07dyMU5nnbPNv6RbN/JXbyDWX7GsRFmwi6djFOx6SjqXbDn0vsG6XdCJC/MnNNXM6M4uAvNZp0rGT3Pv8F0x5+W5qRAbbLbdzXyK5ZgtN65ff9b5/+3fIEQKO3lBdaRxpGU4FfgaCReQ1YC0wobQVK6X2AtFYk+IbIvJKaddZEtENq3MgIZX4xDSyc3KZvSSOXh2bXHnBaySeuF3x1K4WTLWIQFyNBgZ0i2bRavvzSwF+XrbDl6eH92DGvA128wd2j+HnxZtKHEPzBtU4eDiV+KTjZOfk8uvSOHp0sD8v2aN9I2Yu/AuAeSu20D6mDiJCxumzDB31KS891o+WTWsVWvfsJbHc2q1cTnHb/Nu/Q45ycRGHhspyxZahUuobEYkFbsqbdLtSakdpKxaRCOCEUmq6iKQDDxQokiMirnn3CZUbo9HAxDGDGDhyKmazYmi/1jSoXXlX3So6HrPZwpiJM/n5g8cxGIQZczew5+BRXni4N1t2J7Bo9Xbax9Thlcf7oRSs27yf0RNn2paPCg+gaqg/f8TtL3EMRqOBt0bdxuCnPsJssXBnn9bUrxXOhM8W0LRBNXp2aMyQvm0Y8dq3tLrtdUy+nnw6bjgAX/60hn+OpDF52m9MnmY90/Lje48RHOADwNxlm/lu8iMljs3R+P/N3yFHVPYhsCNEqSsfZYpIE6A91kPlP5RSpb40JSI9gElYD7lzgEeBt4FnlVKb8g6f+wFxlzpveEFMTAv1x58lb5X8G/jf4Fxvx0txsrfjVTHqp1KL0q5VC2JjN5UqlQXUbKi6verYtdCZw6Njy+AJlGK7YstQRP6D9YbGX7Ae+n8nIjOUUm+VpmKl1GKs9wLl1ynf/OeA59A07Zrg5A1Dhy6g3AM0V0qdBRCR8cBmoFTJUNO0fxdn79zVkWSYXKCcMW+apmmaQ6xXkys7iqIV1VHDu1jPEZ4AdorI4rzx7hTxsLOmaVohUrlXih1RVMvwwhXjncCCfNM3XKKspmlaka7aw2Sl1JcVGYimadeuq/ow+QIRqQ2MBxoC7hemK6XqlmNcmqZdY5y9ZejIDVZfAf/Dmtx7ATOBH8sxJk3TrkHi4FBZHEmGnnn3BKKUOqCUeglrUtQ0TXOICBhcxKGhsjhya835vI4aDojII1h7lvUp37A0TbvWOPthsiPJ8GnACxiJ9dyhH3BfeQaladq1x8lzoUMdNfyZ9+dpLnbwqmma5jDB4XciV5qibrr+hbw+DC9FKTWgXCLSNO3acxX0WlNUy3BKhUWhlat7X3qsskOw88bSfZUdgp0XulxX2SEU4lHFUNkhlLmr9pyhUmpZRQaiadq1SwDD1ZoMNU3TytJV/wSKpmlaWbhmkqGIuCmlzpdnMJqmXZus3f47dzZ05L3JLUVkO7Avb7ypiHxY7pFpmnZNKcuXyJdLfA6U+QDoAxwHUEptxfpSeU3TNIddC+9NdlFKxRdo4prLKR5N065BAhid/DDZkWR4WERaAkpEDMATwN7yDUvTtGuNk+dCh5Lho1gPlasBKcDSvGmapmkOEbmKH8e7QCl1DBhcAbFomnYNc/Jc6FBP159ziWeUlVIPlUtEmqZdk66F+wyX5vvbHbgVOFw+4Wiadi0SqNSOWx3hyGGyXRf/IvItsLbcItI07dpTyfcQOqIkj+PVBELLOpDKtHTdLl6Y/BNmi4W7+7fl6eHd/1Xx1A/x4pbGYbggbEg4yfJ9xy9Zrkm4D8NbRvHOqoMcSc/CIHB7swiiTO4oBb9sP8qB42dLHU/8vnhWL1iNUoqGMQ1p0bHFJcvt37mfRT8sYtAjgwitav1Kph1NY8XcFWRnZSMiDHpkEEbXkj11umLDbl5+bzYWi4U7+7bmibu72c0/n53LyHHT2f73Yfz9vPjk9WFEhQfa5h85eoJOd73FqPt68eiQLrbpZrOFnve/TXiwH99MerhEsV2Js32nwdqnoTNz5AmUkyJyIm9IB5YAL5RF5SLylIh4XmbecBEp927EzGYLoyfOZNb7j7Fh5kv8/Hssew4ml3e1ThOPAAOahPPZ+gQmLN9PdFU/Qn2qFCrnZnShQ60A4k9cTHata/gDMGnFQT5ZF0+/RqGl/rpbLBZWzltJv3v6MfSJoezdtpcTx04UKpd9Pput67cSGnnxd9litvD7T7/TqW8nho4cyq3334qLwZHnCgozmy28OHkWMyY/zMoZLzBnaRx7/zlqV+b7+esx+XiwbubLPHhHJ974aJ7d/Nc+/JUurRsWWvcXs1ZRp0b5tSec7TsNF18VetU+gSLWO62bAsF5g79SqpZSamYZ1f8UcMlkWFFidx6iVlQQNSKDqOJqZEC3aBau2vaviaeavwdpZ7I5cd2a+KcAACAASURBVDYHs4LNiRk0Civ8ipte9YNZvv84OZaL19JCfdzYl3oGgMxsM+dyLESZ3AstWxwpR1IwBZrwC/DDYDRQt3FdDu4+WKjchmUbiO4QjdF4sdWXcCCBoNAggsODAfDw9MDFpWTJcPPueGpEBlO9qnU/9O8azeI12+3KLF6zg9tvbglAn05NWRu7F6Ws22fR6m1EhQdSt2aY3TJJx9JZtm4nQ/q2KVFcjnC27/QFZZkMRaSniPwtIvtF5Pkiyg0UESUilz68yB9fUTOVdc8uVEqZ84bL9nx9JSLiJSILRGSriOwQkbFABLBCRFbklblXRPaKyF9Au5LWVRzJqRlUDfW3jUeE+pOcmlERVTtFPH7uRtLP5djG08/l4ufualemqp87Jg9Xdqdk2k1Pysji+jAfXAQCPF2JMlnLlcaZU2fw9vO2jXv7eZN52r7eY0nHyMzIpGa9mnbT09PSQWDO13P44aMfiF0TW+I4jqZmEBFiso2Hh5gK7YejqelEhFj3ldFowNfLnRMZZzhz9jwfTV/GqPt6Flrv2Pdn89Jj/cv1njtn+05fICIODQ6sxwBMxfqWzobAnSJSqAkuIj7Ak8CfBeddiiM/m1tEpLkjK7uCnkCSUqqpUqoR8B6QBHRWSnUWkXDgNaxJsD3WD3lJIvKQiGwSkU2paallEJp2OQL0bxTKnB0pheb9lZBORlYOT99Yi1sahXHoxFksJf65dIyyKNYuWkv7nu0LzbNYLCTHJ9P9tu4MfGAgB3cf5PCBir/x4e1pi3jwjk54ebrZTV/yxw6C/L1pUj+qwmOqbNZXhTo2OKAlsF8pdVAplQ38APS/RLlxwAQgy5GVFvUOFKNSKhdoDmwUkQPAGaz/P5RSKtqhsC/aDkwWkQnAfKXUmgK/Aq2AlUqp1Lz6fwTqXmpFSqnPgM8AYmJalOq/X3iwH4kpJ23jSSknCQ/2K80qS6Wi48nIyrVrzZk8jGRkXWwpuhldCPNx4/H21QHwcTNyf6sovvzzMEfSs+yS5BMdapB6pnS9vHn5epGZcbElmJmRibfPxZZidnY2x48dZ/a02QCczTzLghkL6D20N96+3kTUiMDDywOA6nWqk5qcSlTt4iefsGA/ko6l28aTj6UX2g9hwSaSjp0kIsREbq6ZU2eyCPDzYvPOeBas2MobH83lVOY5XERwq2LkaGoGv6/dwbL1uzmfncPpM1mMeO0bpoy9p9jxFcXZvtMXFKM1HCQim/KNf5b3f/6Cqtjf3ncEa/6wEZFoIEoptUBERjtSaVGX2f4CooF+jqzoSpRSe/MCvBl4Q0Sc4rUC0Q2rcyAhlfjENMJDTMxeEsfn44b/a+I5nH6OYK8qBHi6knEuh+ZV/fg2NtE2PyvXwiu/XXwU/bF21Zm7M4Uj6Vm4GqzXB7PNirrBXlgsipTT2aWKJ7RqKOnH08k4mYG3jzd7t++lx+09bPPd3N148IUHbeOzv5xNu57tCK0ail+AH3Fr48jJzsFgMJB4KJFmbZuVKI5m9avxz5FUEpKOExbsx5xlcUwtkLS6t2/ErIV/0aJRTeav3Er7mDqICL9+/KStzNtfLsLLw437busIwIuP9gVgXdw+Pvl+eZknQnC+7zRcvIDioDSl1BXP8V22Lut73t8BhhdnuaKSoQAopQ6UNCi7lYlEACeUUtPzrko/gPX1oz5AGtbj+vdFJBA4BdwObC2LuotiNBqYOGYQA0dOxWxWDO3Xmga1w8u7WqeJx6Jg9rajPNSmGi4i/JWQTsrp8/SsH8zh9HPsPJp52WW9qxh5uG01lIKMrBy+i0sqdTwuBhdu7HMjc7+ei8VioWF0QwJDA9mwbAMhESHUalDrssu6e7jTrG0zZn4yEwRq1K1R6Lyio4xGA+OfHsiQZz7GbLYwuE9r6tUKZ+LnC2laP4oeHRpzZ5/WjBw3nbaDxmHy9eTj14aV9GOXKWf7Tl9QhqdJE4H8zf3IvGkX+ACNgJV5R59hwFwR6aeUyt/itI/vctdEROQI1ux6SUqpy867zPp6AJMAC5CDtbOHNsAIrOcSO4vIvVhv20kHtgDZSqkRRa03JqaF+uPPy34+DXhmzq7KDsGOt5tzvflNvx2vaO1atSA2dlOpUlm1+k3Uc1/OdajsiPY1Y4tqGYqIEWvPWV2xJsGNwBCl1M7LlF8JPFtUIoSiW4YGwBvK5k5JpdRiYHGByZuAD/OV+R/wv7KoT9M05yGUXctQKZUrIiOw5hMDME0ptVNEXgc2KaUcy7oFFJUMk5VSr5dkpZqmaXYEjGV4R7VSaiGwsMC0Vy5TtpMj67ziOUNN07TSKsuWYXkpKhl2rbAoNE275l21nbsqpQo/EKppmlZCTp4L9UvkNU0rf4Jjj7tVJp0MNU0rf3IVHyZrmqaVFesTKDoZapqmOf3tKToZappWIZy8YaiToaZpFcGxvgork06GmqaVO301WdM0LY++gFLOzmab2RqffuWCFaRpddOVC1WwiX3qV3YIdrJzLZUdgp2bJq+u7BAKWTqqY2WHYFMmvZcL+jBZ0zRNHyZrmqbl0S1DTdM09H2GmqZpCGDQLUNN0zR907WmaRogiJMfKOtkqGlahdAtQ03T/vWst9Y4dzbUyVDTtPInumWoaZoG6MfxNE3T8jp3rewoiqaToaZpFUJfTdY0TUOfM3QKf27ey5T/LcRssdC7awxDb73Rbv7MeX+wYNkmDC4umHy9GPP4rYQF+3M09SQvT/wOi1KYcy3c2qs1/Xu0LPd4l67bxQuTf8JssXB3/7Y8Pbx7mdexbP0u/vPubMwWC3f1a8OT93Szm38+O4fHX5vO1r8PE+DrxedvDKdaRCAnMs5w3wtfsnl3AoN7t2LCs7cDkHkmiz6PvG9bPvlYOrf1bMH4pwc6FM+KDbt55f3ZWCyKO/u0ZsTdNxWIJ5cn35jO9r+P4O/rycevDyMqPNA2P/HoSTrd/Raj7u3JI0O6sD8hhUdf+do2PyHpOM8+0IsHB3Uq7qYCoFWtAJ66qQ4uLjBvSzLTNyQUKtOlfjD3dagJCvYdy+S1ubsACPV14/mb6xPi44YCnp25jaMZWcWOYcWG3bz83mwsFgt39m3NE3cX3Ge5jBw3ne1/H8bfz4tPCmyjI0dP0Omutxh1Xy8eHdLFNt1sttDz/rcJD/bjm0kPFzsuR+mWYSUzmy28/8U83n7lXoIDfHnk+U9o16IBNaJCbGXq1Azn0wmP4u5WhTmL/+TTbxcz9pnBBJp8mPrmw1RxNXL23HnufeZD2t1Qn6AA33KNd/TEmfwyZQQRoSa6DJtEr46NqV8rvEzreP7tWcz64HEiQkx0v/dtenZoRL2aF+uYMXcDJl9PNv70Cr8sieX1qXP5Yvy9uFUx8vxDvdlzMJndB5Nt5b293Fn57XO28a7DJtK7U1OH4/nPOz/x/buPEh5i4uYH3qF7+0bUrRlmK/P9/A34+Xjyx48vMWdpHOM/nscnrw+3zX91yq90btXANn5dtVCWfDXGtv6YW8fSq2OTYm8rsJ7rGtW9Lk/9sIVjp87zxfAWrN2XxqHjZ21lIv09uLtNdR79No7TWbmYPF1t817q04Bv1sWz8dBJPFwNWFTx+8Qymy28OHkWP7z3WN42mkyP9o0LbKP1mHw8WDfzZX5dGscbH83j03HDbfNf+/BXurRuWGjdX8xaRZ0aoWSeKX6CdtTVcM7Q2XvVKbU9+49QNSyQiNAAXF2NdGnXmD827rYr07xRLdzdqgDQsE4UqcdPAeDqaqSKq/X3IifXjCrBl7i4YnceolZUEDUig6jiamRAt2gWrtpWpnXE7YqnRmQwNapa67ilWzSLVm+3K7NozXbuuNnaCu7buRlrNu1FKYWXhxutm9XGrYrrpVYNwIGEY6SdzKRNs9oOxbN5dzw1IoOonhdP/5uas3itfTy/r93O7b1uAKB3p6asjd1n2x+/rd5GtfAA6uVLDPmtjd1L9apBRIYFOBRPQQ0ifDly8hxJ6VnkWhTLdqfQoW6QXZl+zSKYHZfI6axcANLP5gBQI9ATg4uw8dBJAM7lmDlfgv4crdso+OI26hrN4jX222jxmh3cnrfP+nRqytrYvbZttGj1NqLCA+2SJ0DSsXSWrdvJkL5tih1TsYjg4uBQWco9GYpIDRHZkW/8WRF5VURWisj7IrJFRHaISMu8+TfmTdsiIptFxKc09aeeOEVwkJ9tPDjQl9QTpy5bfsHyWFo2r2MbP5aWzn3PfMighydxZ/8O5doqBEhOzaBqqL9tPCLUn+TUjDKuI52qIRc7oY0IMRWq42hqBlVDrWWMRgO+3u6cyDjj0Pp/WRLLLTdFO9xl09HUDCJCLn7m8GATRy8Rz4UyRqMBXy93Tmac4czZ80ydsYxn7u152fXPWRrHLTdFOxTLpQR7u3Hs1MVW07HT5wn2cbMrExXgQVSAJx/fHc1n90TTqlZA3nRPMs/n8uaARvzv3hY83rl2iVpI1s9/cZ+FX3KfpRfaRifyttFH05cx6r7C22js+7N56bH+FZKExMGhslR2y9BTKdUMeAyYljftWeDxvOkdgHMFFxKRh0Rkk4hsSj+RVmbB/L56C38fSGRw/w62aSFBJqa98wQzpjzN4lWbOZGeWWb1Xat+WRLHgO4lTz7FMXnabzw4qBNenm6XnJ+dk8vvf+ykT+dm5RqHwUWI9PdgxIzNjJ2zi+d61cPbzYjBRWgaaWLK8v088FUsESZ3bm5cdqc8HPH2tEU8eEfhbbTkjx0E+XvTpH5Uucdw4b3JztwyrOxzht8DKKVWi4iviJiAP4B3RGQGMFspdaTgQkqpz4DPABo0bl7ksWtwgC+paRd/QVOPnyL4Eq27Tdv2M/3nVbz/+v22Q+P8ggJ8qRkVyrbdh+jUplGxPmRxhAf7kZhy0jaelHKS8GC/IpYoSR0mEo9dfFVC0rH0QnWEBfuRmGJtaeTmmjmVmUWAn9cV171jXyK5ZgtN61dzOJ6wYD+Sjl38zMmp6YRdIp6kYyeJCDFZ4zmThb+fF5t3xbNg5RbGfzyXU5nncBEX3NxcuXeg9QdtxYbdNK4bSXBAyQ8wUjPPE+LrbhsP8XEj9fR5+zKnzrMz6RRmiyI5I4vDJ84RGeBB6unz7DuWSVK6tWW5el8a10f4QjHPfFg//8V9lnzJfWYqtI0C/LzYvDOeBSu28sZHF7aR4FbFyNHUDH5fu4Nl63dzPjuH02eyGPHaN0wZe08xt5BjnPyUYYUkw1zsW6Du+f4umMiUUuq/IrIAuBn4Q0R6KKX2lLTyetdV5UjycZJTThAU4MvyP7bz0lO325XZdzCJdz6dw8SXhuHv522bfux4Bn7enri5uXI68xzb98Rze5+2JQ3FIdENq3MgIZX4xDTCQ0zMXhLH5+OGl2kdzRtU45/DqcQnHSc82I9fl8TxyevD7Mr07NCIHxf+xQ2NazJvxRbat6jj0GHv7N9jGdA9pljxNKtfjX8Op5GQdJywYD/mLN3M1LF325Xp3q4RsxZtpEWjmixYuZV20dZ4fvlopK3M5C8X4eXhZkuEAL+W8hAZYE/SaSL9PQj3cyf19Hm6Ngjltbk77cqs3pdGt4YhLNx+FD8PV6ICPEhKP0dmVi7ebkZMHq6kn8shpro/e5JPFzuGZvWr8c+R1IvbaFkcUwskre7tGzFr4V+0aFST+Su30j7Guo1+/fhJW5m387bRfbdZ37Hy4qN9AVgXt49Pvl9ebokQcPpsWBHJMAUIEZFAIBPoA/yWN+8OYIWItAcylFIZIlJbKbUd2C4iNwD1gRInQ6PBwJMP9GH0G19jsVjo1SWGmlGhTPthKfVqV6XdDQ34+NvfOJeVzdjJPwAQGmTizefvIuFIKh99vQgRQSnFHf3aU6v6pU/SlxWj0cDEMYMYOHIqZrNiaL/WNKhdtodVRqOBt569jUFPfmS9TaNPa+rXCue/ny2gWf1q9OzYmKF92/DYa99yw22v4+/ryWfjhtuWj77lVU6fzSI7J5dFq7Yx64PHbFei5y7bzPfvPFLseN54ZiBDnvkEi8XCHb1bUa9WOJO+WEjT+tXo3r4Rg/u0ZuS46bS74w1Mvp589OqV/9OePXee1Rv/ZsLoQcWKpyCzUry7ZC/vDG6KQYT525L5J+0sD3SoyZ7kU6zdf5w/D56gZc0Apj/YEotFMXX5AU6ds15Mmbp8P+8PaYYAfx89zdwtScWOwWg0MP7pgQx55mPMZguD+7SmXq1wJn6+kKb1o+jRoTF35m2jtoPGYfL15OPXhl15xRXI2R/Hk4q4QioiI4EngUTgIHAI6ARsAW4EXIH7lFJ/iciHQGfAAuwEhiulzl9itYD1MPmrX1eUa/zF4Yxvx8s1O9fb6Jzt7Xjd3llT2SEU4kxvx+vSvhWb4zaVKpM1aNxcfTNnpUNlW9Y2xSqlWpSmvpKokHOGSqkPgA/yTxORlcB0pdRTBco+URExaZpWwZy7YVjpF1A0TfsXsN4249zZsNJurVFKdVJKbaqs+jVNq0B5/Rk6Mji0OpGeIvK3iOwXkecvMf8ZEdklIttEZJmIVL/SOiv7PkNN0/4lyuqmaxExAFOBXkBD4E4RKfic4WaghVKqCfATMPFK69XJUNO0CiCIODY4oCWwXyl1UCmVDfwA9M9fQCm1Qil14eHxDUDklVaqk6GmaRWiGIfJQReeMMsbHiqwqqrA4XzjR/KmXc79wKIrxacvoGiaVu6K+dxxWlndWiMidwEtsN7CVySdDDVNqxhldzE5Ecj/QHVk3jT76kRuAv4D3FjUvcoX6MNkTdMqhDj4zwEbgToiUlNEqgCDgbl2dYk0Bz4F+imljjmyUt0y1DStQpTV03hKqVwRGQEsBgzANKXUThF5HdiklJoLTAK8gVl5F2USlFL9ilqvToaappW/Mn5vslJqIbCwwLRX8v19U6GFrkAnQ03TKoSzP4Gik6GmaeVO0G/H0zRNA5y+n4arPxm6uxqoF1Gq16Rc8w6lnr1yoQpUM+TKPWZXpLXPd6rsEAqp/cQvlR2CTdrh9CsXcoSTZ8OrPhlqmnZ1cPbOXXUy1DStQjh3KtTJUNO0iuLk2VAnQ03Tyt3V0LmrToaappW/Mr7pujzoZKhpWoVw8lyok6GmaRXB4Y5bK41OhpqmVQgnz4U6GWqaVv6K2blrpdDJUNO0iuHk2VAnQ03TKoS+tUbTNA19zlDTNA0EXHQyrBzLN+zmlfdmYzZbGNK3NU/c081u/vnsXEaOm862PYfx9/Pi03HDiAoPZPOueEZP+BEApRSj7u/JzTc2BSDj9FlGvfUDew4mIyK8++KdtGhcs8xjX7puFy9M/gmzxcLd/dvy9PDuZV5Hfuti/2by5/OwWBT9u93A8Ns72c2P23GQdz6fz/5DRxk/5k66tmtsm/fB/xayduMelFK0alaHUQ/1LdEtFMvW7+LFd37GYrFwV782PDnM/jOfz87hsde+te2vL964l2oRgZzIOMO9z3/Jlt3xDO7digmjBxVa99BnPyU+8Thrv3+xyBiWrt/Fi5N/ztvubXjqEjE8+uq3bM2LYdp4awwA7371O9Pnrsfg4sJbo26ja5sG7ItP4f4X/2db/lDScV546GYevbMz//1sId/OWUegyRuAlx/rS7d21zu8vTo2COWV25rg4iLMXHeIT5bstZv/0oDGtK4bDIBHFQOB3m40GzOf1nWCeGlgE1u52qE+jPzfXyzZluxw3SXn3NnQKZKhiBzC+jo/b2C+UqpRadZnNlt48e1Z/Pj+Y4SHmOh1/2S6d2hMvZphtjLfz1uPn48H62e9zK9L4njjo3l8Om449WqF89uXozAaDaSkZdD1nol0b9cIo9HAy+/NpnPrBnzx5n1k5+RyLiu7NGFeNvbRE2fyy5QRRISa6DJsEr06NqZ+rfAyr+tCfRM/mcOUcfcTGujHsGem0LFVA2pVC7WVCQs2Mfap25n+y2q7Zbfujmfr7ni+//ApAB587hPidhwkpnHtYsfw3KRZ/PTh40SEmOg2fBI9OzSmXr7PPGPuekw+nmz8eSyzf4/ltalz+HL8fbhVMfLCw73ZfTCZPQeSCq17/ooteHm4ORTDmImzmD3FGkPXYdYY8m/36XkxxM4ey8+/x/LqlDlMe/M+9hxMZvbvsaz74UWOpmZw64ipbPzpZepUD2X1jOdt67++90v06dTUtr5H7uzME3d1Lda2AmsL67VBTblnylqOpp/j19GdWbo9mf1HT9vKvDF7u+3ve26sxfWRJgA27Eujz3+XA+Dn6cqKsT1Ys9uh9yWVytXQues1+Xa8zbviqREZTPWqQVRxNdL/pmgWr9luV+a3NTsY1KslAH06N2XNpr0opfB0r4LRaACsrccLO/BU5jk2bDnAkL6tAajiasTPx7PMY4/deYhaUUHUiLTGPqBbNAtXbSvzei7Yue8wUeGBRIYF4upqpFvHpqz6c5ddmYjQAOrUDC/U4hOB7OxccnLN5OTkkms2E2Aqft+ScbviqRkZRI28/XVrtxgWrbbfX4tWb2dw71YA9OvSjDUbrfvLy8ON1s1q416l8O965tnzfPzdCkbd2+OKMcTutI9hQPfCMSxcdTGG/l2asTovhkWrtzOgewxuVVypXjWImpFBxO6Mt1t21ca/qREZRFR4QLG2zaU0rRFAfNoZDh8/S45ZMT/uCN2aXP7Hsm9MFPNijxSa3qt5VVbtOkpWjrnUMTlCHBwqS4UkQxEZLSIj8/5+V0SW5/3dRURmXGaZWiKyWURuKG59R1MzqBpqso2HB5s4mppRoEw6EaH+ABiNBny93DmRcQaAuJ2HuHHoW3S++79MGDMIo9FAQtJxAk3ePDX+O7oNm8iot77n7Lkrvoq12JJTM6iaFxdARKg/yQViL0upx08RGuRnGw8N9CP1+CmHlm1SvzoxjWvRa9h4eg4bT+vmdakZFVLsGJKPXdwXABEhJpJT7TsUTU7NoGqIdZ8ajQZ8vT1s++ty3vp0Po8N7YKHe5Urx5Cabr/dLxdDaOEYHFl29pI4BnaPsZv2xazVtB/yFiPGzSD9lOMd8Ib5uZN88tzFuE6eI9TP45JlI/w9iAr0Yt3fhVt/faIjL5kky4uIY0NlqaiW4RqgQ97fLQBvEXHNm7a6YGERqQf8DAxXSm28xPyHRGSTiGxKS0st82Cjr6/BqhkvsOjLUXz4zVKyzueQa7awfe8Rht3ajiVfj8HDvQoffru0zOu+mhxOSuPQkWMs+N8LLPzqRTZtO8Dmnf9UdlgAbN97hEOJafTOd1haWbJzcvlt9Xb6d21um3bfwPbEzR7L6unPERboy0vvl0/P1n1joli0JRGLsp8e7OtOvQg/Vu9KKZd6L0VEHBoqS0Ulw1ggRkR8gfPAeqxJsQPWRJlfMDAHGKqU2nqplSmlPlNKtVBKtQgKCi40PyzYj8SUi7/MyanphAX7FShjIinlJAC5uWZOnckiwM++O/q6NcLw8nBjz8FkIkJMhAebiL6+BgB9Ojdj+99l/6saHuxHYl5cAEkpJwkvEHtZCg70JSXtYssz5XgGwYG+Di27csNOGtWrhqeHG54ebrSJqcv2PQnFjiE85OK+AEg6lk54sMm+TLAfices+zQ318ypzHOF9ld+G7f/w5bdCTS/ZSy9H3qPAwnH6Pfo+5ePIdhkv90vF0NK4RiutOzSdbtoUj+KkHzbNSTQF4PBBRcXF+65pS1xBQ6ri3I0I4tw/4stwXB/D1Iyzl2ybJ+YSOZtOlxoeu/oqvy+LYncglmyHOnDZEAplQP8AwwH1mFNgJ2B64DdBYpnAAlA+5LW16xBNf45kkpC0nGyc3KZszSOHu3tr8n06NCImYv+AmD+iq20j6mDiJCQdJzcXOs5lMPJJ9ifkEJUeAAhgb5EhJrYH2/9JV27aS91812QKSvRDatzICGV+MQ0snNymb0kjl4dm1x5wRJqWCeShKTjJB49QU5OLktWb6Vjy4YOLRsabCJuxz/kms3k5pqJ2/EPNaIK/zhdSfMG1Th4OJX4JOtn/mVJLD07NrYr07NDY35Y8CcAc5dvoUOLukW2Iu4b2IGdC8az+dfXWPDZU9SuFsLcj5+8bPnohnkxXNjuv8fSs4N9DL06XoxhTr4YenZozOzfYzmfnUN8YhoHD6cSc31123I//x5b6BD5aL4foPkrt9KgtuMXyLbFn6RGsDeRgZ64GoQ+0ZEsvcTV4Fqh3vh5uhL3z4lC8/rGRF0ySZYXRw+RK/MwuSKvJq8BngXuA7YD7wCxSilV4EudDdwKLBaRTKXUd8WtyGg08OYzA7nz6Y8xmy0M7tOaerXCmfj5QprWj6JHh8bc2ac1T7w+nTa3j8Pk68knrw8D4M+tB5kyfSmuRgMiwlujbrfd/jD+6YE8/tq35OTkUi0iiPf+M6RUG+RysU8cM4iBI6diNiuG9mtdrP8oxa7PYGDMI/0YOXYaZouFfje1oHb1UD6Z/jsN6kRyY6uG7Nx7mDFvfsupzHOs3biHT2csYeZHz9C1bWM2bT3AnSPeQ0RoE13X4URqF4PRwH+fvZ3bR36ExaIY0rc19WuF89anC2jWoBq9OjZmaL82PPbqN9ww8DVMvp58/sa9tuWb3zKW02eyyMnJZeGq7fz0wWN2V6IdjWHi6Nu5beRHmC2KoX2t2/3NTxfQPC+Gu/q14ZGx3xAz4DX8fT35Yrw1hga1w7nlpmja3PEmRoMLE8fcjsFgbWecOXeelX/u4d0XBtvV9+qHc9i+9wgiQrXwAN4pML8oZovi1Zlb+PrxdriIMGtDPPuOnuap3g3YnpDOsu3WxNg3Jor5lzgnWDXAk3B/D/7cn1asbVRazv4EiihVMc1kEekK/AaYlFJnRGQv8IlS6p1L3VojIiZgCTBOKTX3cuttHt1Crfjjzwr4BI5xdzVUdgiF7D+aWdkh2HG2t+M5483ATvV2vJ/HkJ26v1RbjNeO9wAACWRJREFUqVl0jFqy2rH/pyE+rrFKqRalqa8kKqxlqJRaBrjmG6+b7+8aeX+mAY3ypqUDxb6SrGmac3LC3xw7TnHTtaZp1zrRrwrVNE3TT6BomqZdJXTLUNO0CuHsLUOdDDVNqxDOfmuNToaappU//d5kTdO0q+MCik6GmqZVCH2YrGmahvO3DPWtNZqmVYiy7LVGRHqKyN8isl9Enr/EfDcR+fH/7Z17kNVlGcc/X1AERMFRsbzkKqBWFox4SREVNfJCSSVDpmMoalBWTEHjpJmWkzA0OTnlFJljZWNm5uRo5jAKsiJyW1jBG3jLSVOxbAuQ5PLtj/ddOnvcs7vQOXvOcp7PzM7+Lu/+nu85++5z3vfd3/P95fOLJTV0ds1IhkEQdA9lyoaSegM/Ac4GPgRcIKnYIWQy8LbtocBNwKzOrhvJMAiCiiOgl9Slry5wPPC87Rdtvwv8FjivqM15wC/z9u+BM9SJc2yPXzNcuWL5W/v0363rzpil2Y9kFFEr1JoeqD1NoadjyqXn0M6bdExT0/KH+u2u/brYvK+kZQX7c2zPKdg/CCg0Y/wrcELRNba3sb1FUguwLx28Hz0+GdrecTfRdpC0rBq2QaWoNT1Qe5pCT8fUkh7bZ1VbQ2fENDkIgp7Gq8AhBfsH52PttpG0GzAQ+HtHF41kGARBT2MpMEzSYZL6AJ8Dig2g7wO+kLfPBx5xJ07WPX6aXEbmdN6kW6k1PVB7mkJPx9SanrKQ1wCvBB4CegO32X5K0neBZdkZ/xfAryU9D/yDlDA7pNts/4MgCGqZmCYHQRAQyTAIggCo02Qo6TpJ06utoxVJj1dbQ2dIurX1Ln9J36qylmmS+pc4N0nSj7tbUy0i6WVJ+0lqkLS62npqnbpMhrWG7ZOqraEzbF9m++m8W9VkCEwD2k2GQbCz1E0ylHS1pDWSHgOOBHpJWp7PDZdkSR/I+y9I6i9pgqTVkpolLaigtvVKzM7xVkmamM/1knSLpGclzZX0J0nnV0pLjrmnpAfy614taaKk+ZKOlTQT6CdppaTfVFJHCS3fAQ4E5kmal9tckn+3S4BRFdTSZoQlaXqeZcyX9KP8nqyWdHw+f2o+tlLSCkl7lVnPDElfzds3SXokb59e6ncj6fCsJR7DW0Rd3FojaSTpX+sjSK+5CVhOKvvZGxgNLANG52T5pu2Nkq4FPmH71fxQ+0rymaxvOKmMamlOwKOABlJB+mDgGeC2Cms5C3jN9rkAkgYCUwFsXyXpStsjKqyhIy2XAGNsvyXp/cD1wEigBZgHrOgmbYX0tz1C0imk38/RwHTgy7YXShoAbCpzzEbgG8DNwLHAHpJ2J/Xn1r6zHUlHkup4J9luLrOWHk+9jAxHA/fa3mj7X/zvBs3HSR3mFOD7+ftoUicDWAjcLuly0v1MleRk4E7bW22/ATwKHJeP3217m+3XSX/slWYV8HFJsySNtt3SDTF3VssJwHzb63LR/l3dLxGAOwFsLwD2zh+eC4Ef5tHbINtbyhxzOTAyf6D/B1hESoqFfbiV/YE/AhdGImyfekmGpVhA6jiHkjrKcFLyaQSwPQW4hlTWs1zSvlXS2a3YXgMcQ0pEN+QRct1rAbbQ9m+mb8F28Q27tj0TuAzoByyUdFQ5xdjeDLwETCJ9sDcCY4ChpBlEIS3AK6T+HbRDvSTDBcB4Sf3yus0n8/FG4CJgre1tpDvVzwEeA5A0xPZi29cC62hbD1luGoGJknpL2p80Sl1CGl18Nq8dHgCcVkENAEg6ENho+w5gNikZFbI5T8cqTgkt/wZa198WA6dK2jdrmlBBOW8Ag3OsPYBxBeda13hPBlpst+T+s8r2LFIJWVmTYaaRNB1fkLenACvaKT17F/g0cLGkz1dAR4+nLtYMbTdJugtoBt4kdUxsvyxJpI4EKQkebPvtvD9b0jCSHdvD+ecrIhG4FzgxxzDwTduvS7oHOAN4mmRJ1ET6lK8kHyG99m3AZtJ64Q8Kzs8BnpTUZPvCKmg5EfizpNdsj5F0HWmK+E9gZaWE2N6sVPK1hGQE8GzB6U2SVgC7A5fmY9MkjQG2AU8BD1ZAViNwNbDI9gZJm3jvFLlV/wZJ44C5ktbnsrUgE+V4VSZPvZtsl/SMkzTA9vrcdgkwKq8fBjWApPnAdNvLOmsb1C51MTKsVfIUcD5tR13tcX9ekO8DfC8SYRCUnxgZBkEQUD//QAmCIOiQSIZBEAREMgyCIAAiGe7ySNpaUDN7t0q4vXTxWqdJuj9vf0rtPLy7oO0gSV/aiRjtOgqVOl7U5vYdqdsurjUO6ptIhrs+79geYfto0o23UwpPKrHD/cD2fbnCohSDgB1OhkFQLSIZ1heNwNA8InpO0q+A1cAhksZKWiSpKY8gBwBIOkvJMaeJZCZBPr7dN1DSAZLuVXKWaZZ0EjATGJJHpbNzuxmSlkp6UtL1BdcqdhTqEEmX5+s0S7qnaLR7pqRl+XrjcvveSo5ArbG/+P++kcGuRyTDOkHpcYlnk2p8AYYBt9j+MLCBVIN9pu1jSA4+X5fUF/g5qXxxJPC+Epe/GXjU9nBSudxTwFXAC3lUOkPS2BzzeJI7z0hJp6ito9A5JHOKzviD7eNyvGeAyQXnGnKMc4Gf5tcwmVQid1y+/uWSDutCnKCOiJuud336SWotUWskPTXsQOAvtp/Ixz9GsghbmKoT6UMqbzsKeMn2WgBJdwBXtBPjdOBiANtbgRZJ+xS1GZu/Wu21BpCS415kR6EcoyslYkdLuoE0FR9AekpaK7/LdeZrJb2YX8NY4KMF64kDc+w1XYgV1AmRDHd93in2HswJb0PhIWCu7QuK2pXTs1DAjbZ/VhRj2k5c63ZgvO1mSZNoa17xHveYHPsrtguTJpIadiJ2sIsS0+QA4AlglKShsN1d+giSEUGDpCG53QUlfv5hsvlrXp8bSFtnGUijt0sL1iIPkjSY0o5CHbEX8LfsUlNsFDFByeFnCHA48FyOPTW3R9IRkvbsQpygjoiRYYDtdXmEdaeSNRXANbbXSLoCeEDSRtI0uz3r+q8BcyRNBrYCU20vkrQw37ryYF43/CCwKI9M1wMXlXIU6oRvk6y71uXvhZpeIZlZ7A1Msb1J0q2ktcQmpeDrgPFde3eCeiFqk4MgCIhpchAEARDJMAiCAIhkGARBAEQyDIIgACIZBkEQAJEMgyAIgEiGQRAEAPwX4LlubmMK+5YAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_confusion_matrix(clf, X_train, y_train,\n",
    "                      display_labels=label_encoder.classes_,\n",
    "                      cmap=plt.cm.Blues,\n",
    "                      normalize=\"true\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "         dws       0.64      0.91      0.75       123\n",
      "         jog       1.00      0.26      0.41       119\n",
      "         sit       0.55      0.98      0.70       341\n",
      "         std       0.80      0.44      0.56       285\n",
      "         ups       0.85      0.64      0.73       138\n",
      "         wlk       0.97      0.79      0.87       346\n",
      "\n",
      "    accuracy                           0.71      1352\n",
      "   macro avg       0.80      0.67      0.67      1352\n",
      "weighted avg       0.79      0.71      0.70      1352\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_test, clf.predict(X_test), target_names=label_encoder.classes_))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<sklearn.metrics._plot.confusion_matrix.ConfusionMatrixDisplay at 0x12bd951d0>"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_confusion_matrix(clf, X_test, y_test,\n",
    "                      display_labels=label_encoder.classes_,\n",
    "                      cmap=plt.cm.Blues,\n",
    "                      normalize=\"true\")"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.5"
  },
  "pycharm": {
   "stem_cell": {
    "cell_type": "raw",
    "metadata": {
     "collapsed": false
    },
    "source": []
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}