language_detection_model.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "6d6c52b5",
   "metadata": {},
   "source": [
    "### This is a simple language detection model with sklearn"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "551abf92",
   "metadata": {},
   "source": [
    "First we import all the neccessary libraries, import the csv file and take a look at the data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 240,
   "id": "0f89c823",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.naive_bayes import MultinomialNB\n",
    "from sklearn.linear_model import SGDClassifier\n",
    "from sklearn.pipeline import Pipeline\n",
    "import pandas as pd\n",
    "from sklearn.metrics import confusion_matrix, classification_report, ConfusionMatrixDisplay\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "from sklearn.feature_extraction.text import CountVectorizer, TfidfVectorizer\n",
    "from sklearn.calibration import CalibratedClassifierCV"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "19768ad0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>text</th>\n",
       "      <th>language</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>ich denke es handelt sich hier um ein missvers...</td>\n",
       "      <td>german</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>ich habe tom gerade erst verlassen</td>\n",
       "      <td>german</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>tom versuchte mary nur zu ärgern</td>\n",
       "      <td>german</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>tom hat mir die hand geküsst</td>\n",
       "      <td>german</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>ich wusste dass dir das gefiele</td>\n",
       "      <td>german</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                text language\n",
       "0  ich denke es handelt sich hier um ein missvers...   german\n",
       "1                 ich habe tom gerade erst verlassen   german\n",
       "2                   tom versuchte mary nur zu ärgern   german\n",
       "3                       tom hat mir die hand geküsst   german\n",
       "4                    ich wusste dass dir das gefiele   german"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = pd.read_csv(\"languages.csv\")\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 157,
   "id": "dec23e3a",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "english    275687\n",
       "german     199618\n",
       "french     169693\n",
       "spanish    118686\n",
       "Name: language, dtype: int64"
      ]
     },
     "execution_count": 157,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.language.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 158,
   "id": "bb13a71b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "text        0\n",
       "language    0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 158,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.isna().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 159,
   "id": "75b58327",
   "metadata": {},
   "outputs": [],
   "source": [
    "plt_y = list(data.language.value_counts())\n",
    "plt_x = list(set(data.language.values))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 160,
   "id": "a55d6794",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<BarContainer object of 4 artists>"
      ]
     },
     "execution_count": 160,
     "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.bar(plt_x, plt_y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "82a9e8d0",
   "metadata": {},
   "outputs": [],
   "source": [
    "## Shuffle the rows since a quick look at the table head lets us see that languages are strictly separated, \n",
    "## which is no good for the model\n",
    "\n",
    "shuffled_data = data.sample(frac=1).reset_index(drop = True)\n",
    "X_raw = shuffled_data[\"text\"]\n",
    "y_raw = shuffled_data[\"language\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 162,
   "id": "694b11e2",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(763684,)\n",
      "(763684,)\n"
     ]
    }
   ],
   "source": [
    "print(X_raw.shape)\n",
    "print(y_raw.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f3e31be8",
   "metadata": {},
   "source": [
    "There are 4 different languages present in the table - english, german, spanish adn french. The data is not uniformaly distributed with french texts being 2.5 times less than for english. For now I do not see it as a problem, so I will leave it as is and separate the data to X (features) and y(target).\n",
    "\n",
    "As the data comes in textual format, I will need to use a CountVectorizer to transform the text to bag-of-words dictionary. For target 4 language I will use LabelEncoder to transform. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "11923a33",
   "metadata": {},
   "outputs": [],
   "source": [
    "le = LabelEncoder()\n",
    "y = le.fit_transform(y_raw)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "4a7d1924",
   "metadata": {},
   "outputs": [],
   "source": [
    "cv = CountVectorizer()\n",
    "X = cv.fit_transform(X_raw)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 246,
   "id": "ed7bc805",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.2, random_state = 1)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5e615589",
   "metadata": {},
   "source": [
    "Now I initiate a Multinomial Bayes for classification, no hyperparameters are set at this point"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 247,
   "id": "b412d475",
   "metadata": {},
   "outputs": [],
   "source": [
    "# clf = MultinomialNB()\n",
    "clf = CalibratedClassifierCV(MultinomialNB(), cv=3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 248,
   "id": "8ec7f3dd",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "CalibratedClassifierCV(base_estimator=MultinomialNB(), cv=3)"
      ]
     },
     "execution_count": 248,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 249,
   "id": "aa1f0e3d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9990113724899664"
      ]
     },
     "execution_count": 249,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 269,
   "id": "b4b669d7",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<sklearn.metrics._plot.confusion_matrix.ConfusionMatrixDisplay at 0x26348e3ef40>"
      ]
     },
     "execution_count": 269,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWUAAAEGCAYAAAC95YRPAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8QVMy6AAAACXBIWXMAAAsTAAALEwEAmpwYAAA1/UlEQVR4nO3deXxU1fnH8c83C0kgEAj7ooCAC26oiFKXqtiira221Uq1P7TlV61LbWtrq21dW38udalWRam2olZFqa1bFXetVkFAZVMEQQFBIGEJS8g2z++PewJDyDIhy8yE5/163VfunLn33Ocy4ZmTc889V2aGc8651JCR7ACcc85t40nZOedSiCdl55xLIZ6UnXMuhXhSds65FJKV7ABSSbfCTBuwW3ayw2h2H89qn+wQnANgC5sotzI1pY7Rx3aw4jVVCW07Y1bZFDM7oSnHa22elOMM2C2baVN2S3YYzW50n2HJDsE5AKbay02uo3hNFdOm7J7Qtpm9F3Rr8gFbmSdl51xaMSBGLNlhtBhPys65tGIYFZZY90U68qTsnEs73lJ2zrkUYRhVbXh6CE/Kzrm0E8OTsnPOpQQDqjwpO+dc6vCWsnPOpQgDKrxP2TnnUoNh3n3hnHMpw6Cq7eZkT8rOufQS3dHXdnlSds6lGVFFk+Y0SmmelJ1zaSW60OdJ2TnnUkI0TtmTsnPOpYyYt5Sdcy41tPWWsj8OyjmXVgxRRUZCSyIkfSpptqT3JU0PZYWSXpS0IPzsErf9ZZIWSpovaXRc+SGhnoWSbpekUJ4jaVIonyppQH3xeFJ2zqWdmCmhpRGONbNhZjY8vL4UeNnMhgAvh9dIGgqMAfYFTgDukpQZ9hkPnAMMCUv1Y6jGAWvNbDBwK3BDfYF4UnbOpRVDlFtmQksTnAxMDOsTgVPiyh81szIzWwwsBEZI6g10MrO3zcyAB2rsU13XZGBUdSu6Np6UnXNpJbp5JCOhBegmaXrcck4dVb4gaUbc+z3NbAVA+NkjlPcFlsbtuyyU9Q3rNcu328fMKoH1QNe6zs8v9Dnn0k4jLvQVxXVJ1OUIM1suqQfwoqSP6tm2tgNbPeX17VMrT8pNNHbEUPLyq8jIgMws447nP+bBm3rx3MOFFBRGzxH7wWXLGTFqAzNez+ev/9eHygqRlW386PLlDDty43b1XXnWQFYsaceEV+dvV/6fZwr4wzkD+fNz89nzwNJWO7/6XHzLEg47fgPrirI497i9ABh7yQpGji7BDNYVZXHTz3ZnzcrsJEfaNB06VfHzm5YyYO8tmMEtF+/GhzM6JDusZvGtH63mxDOKMROLP8rl5p/vRkVZav8BbSaqrPliNLPl4ecqSf8ERgArJfU2sxWha2JV2HwZEP/I+37A8lDer5by+H2WScoCCoA1dcWT2v/6cSQNkDQnrA+XdHs92x4j6ZnWiu3Gxxcy/qX53PH8x1vLvvWj1Yx/aT7jX5rPiFEbACgorOKaiYu455X5XHLbEm68aPvHpL/57wJyO+x4V//mjRn8677u7H3wppY9kUZ6YVIhvz1z4HZlk8f34Lzj9+L8r+zF1Jc68f2fr0xSdM3nvGs+Z/prHfnfo/fmvOP3ZMmC3GSH1Cy69qrglHFFXHjinpx73F5kZhjHnLwu2WElJIYSWhoiqYOkjtXrwFeBOcBTwFlhs7OAJ8P6U8CYMKJiINEFvWmhi2ODpMNDf/HYGvtU13Uq8Erod65V2iTleGY23cwuSnYcjTV4/1K69qoEoP9eWygvy6C8LPrFKd2UwRP3dOeMn32xw34Tb+zNaeevol1Oak2NNWdqPhvWbv/H1uaN2y6u5ObFSPdpb9vnV7H/4Zt4/uFCACorMthU0qQLSCklM8vIyY2RkWnk5MUoToO/aqILfVkJLQnoCbwp6QNgGvCsmT0PXA98RdIC4CvhNWY2F3gMmAc8D1xgtvXR2ucB9xJd/PsEeC6U3wd0lbQQuJgwkqMurdJ9Ien7wEVAO2AqcD5RZ/dtwElAKXCyma2UNAj4O5BJdFIXm1l+jfqOAX5pZidJ+nKoB6J+mqPDer6kycB+wAzg+/V9O+38yRm/+d4gEHz9f4r52veLAXj6b915eXIhQw7YzDlXLqdj5+0fif7mswUM2rd0a6KdeGMvvvPj1eTkbR/iwtl5rF6ezeFfKeEfd/cgHZz96xUcf9paNpVk8qtTByU7nCbp1b+c9cWZ/OLWpeyxbykLZrVn/OV9KCtN/8Rc/EU2k8d358F3P6Rsi5j5ekdmvt4x2WE1qPpCX7PUZbYIOLCW8mJgVB37XAtcW0v5dKJ8U7N8C3BaojG1eEtZ0j7A6USd6cOAKuBMoAPwjpkdCLwB/Cjschtwm5kdyrY+mfr8kujbahhwFFGCBzgI+BkwFNgDOKIZTmcHtz65gDtf+Jhr/76Ip+7vxux3OnDSWUX87e153PXifAp7VjDh6j7b7fPp/Fzuu7YPP70xuoj7yZw8li/O4YgT12+3XSwG91zVl3OuTOSfIXXcf0Nvvj98KK880Zlv/rAo2eE0SWamMXj/Up55oCsXfHUvtmzO4PQLVzW8YxrIL6hk5OgSzjpsH844aF9y28c47ttrkx1WQqpMCS3pqDW6L0YBhwDvSno/vN4DKAeq+31nAAPC+kjg8bD+cAL1vwXcIukioHMYcgJRP88yM4sB78fVvx1J51QPl1ldXFXbJvWq7o7o3K2SI05Yz0fvtadL90oyMyEjA048cw3z32+/dfvVy7O5ZtwALrltCX0GlAMwb0Z7Fsxuz9gRQ/nFKYP5fFEOl3xnMKUbM/j0o1x+9Z3BjB0xlA9ntufKs/fg4w/yGh1nMrz6zy4c+bX1DW+YwopWZLN6RTbz34su7L35TAGD90+NC61NddBRG/liaTvWr8miqlK89e8Chg5PresWtWnuO/pSTWt0XwiYaGaXbVco/TKuO6FqZ2Mxs+slPQt8DXhH0vHhrbK4zeqs38wmABMAhh+Y26jujS2bM4jFoH1+jC2bM5jxekfOvPgLildm0bVnlKz/+1wBA/baAsDG9ZlcPnYPfnDZCvYdse2X/xtnFfONs6Jujy+WtuOKsQP54z8WAvD43Dlbt7vkO4P50RWfp8zoi9r0GVjG8sU5ABw+ej1LF+YkOaKmWbs6m6Ll7eg3aAvLPsll2FEb28yFvlWfZ7PPwZvIyYtRViqGHbmRj2elxxd+rBlHX6Sa1kjKLwNPSro1DDkpBOrruHoH+A4wieh2xnpJGmRms4HZkkYCewPrmh52w9auzuLqcdHog6pKOPZb6zj02A3c+JPd+WRuHhL07FfORaGb4qm/dWP54nY8fGsvHr61FwDXPfoJnbtV1nmMVHbpXZ9xwMiNFBRW8tD0eTx4c09GHLeBfoPKiMVg1eftuP3X/RquKMXd+bu+/PqOJWRlG18sacfNP9+t4Z3SwPz3OvCfZztz55SPqaoUC+fk8dxDdd7TkDKiCYnablJWS1z72uEg0unAZUTdJRXABcBL1RfwJJ0KnGRmZ0saAjxE1MJ+FjjHzPqGSTyeMbP9alzo+zNwLFFreB5wNlEXyC/N7KRQ/x3AdDO7v744hx+Ya9OmtI3/cPFG9xmW7BCcA2CqvUyJrWlSZ+/A/fPtqicOSGjbs/d8e0YCN4+klFYZfWFmk4havvHy496fTHRPOMDnwOFmZpLGANPDNp8Srmya2WvAa2H9J7Uccuv7YZsLm3wSzrmUYEaz3jySalLxjr5DgDvCAOx1wA+TG45zLrUkdmNIukq5pGxm/6GWcYPOOQehT9lbys45lzra8oU+T8rOubRiNHoC+7TiSdk5l1YMqEhsXou01HbPzDnXRqlNPzjVk7JzLq0Yfkefc86lFG8pO+dcijCTt5Sdcy5VRBf60n8+67p4UnbOpZnmfUZfqvGk7JxLK9GFPu9Tds65lOF39DnnXIrwO/qccy7FNNeDU1ORJ2XnXFoxg4qYJ2XnnEsJUfeFJ2XnnEsZfkefc86lCB8S55xzKcW7L5xzLqX4M/p2ER/Pas/oPsOSHUaz++yx/ZMdQovp/93ZyQ7BtbJo9IXPfeGccynBbx5xzrkU05a7L9pub7lzrk2qHn2RyJIISZmS3pP0THhdKOlFSQvCzy5x214maaGk+ZJGx5UfIml2eO92SQrlOZImhfKpkgY0FI8nZedc2olZRkJLgn4KfBj3+lLgZTMbArwcXiNpKDAG2Bc4AbhLUnXn9njgHGBIWE4I5eOAtWY2GLgVuKGhYDwpO+fSipmotIyEloZI6gd8Hbg3rvhkYGJYnwicElf+qJmVmdliYCEwQlJvoJOZvW1mBjxQY5/quiYDo6pb0XXxPmXnXNppxIW+bpKmx72eYGYT4l7/CfgV0DGurKeZrQAwsxWSeoTyvsA7cdstC2UVYb1mefU+S0NdlZLWA12BoroC9qTsnEsrjbyjr8jMhtf2hqSTgFVmNkPSMQnUVdtBrZ7y+vapkydl51zaaaYhcUcA35T0NSAX6CTpIWClpN6hldwbWBW2XwbsFrd/P2B5KO9XS3n8PsskZQEFwJr6gvI+ZedcWqkep9zU0RdmdpmZ9TOzAUQX8F4xs+8DTwFnhc3OAp4M608BY8KIioFEF/Smha6ODZIOD/3FY2vsU13XqeEY3lJ2zrUtLTxO+XrgMUnjgCXAaQBmNlfSY8A8oBK4wMyqwj7nAfcDecBzYQG4D3hQ0kKiFvKYhg7uSdk5l1bMoLKZJ7k3s9eA18J6MTCqju2uBa6tpXw6sF8t5VsIST1RnpSdc2nHb7N2zrkU4XNfOOdcijFPys45lzra8oREnpSdc2nFzPuUnXMuhYiqZh59kUo8KTvn0o73KTvnXIrwp1k751wqsahfua3ypOycSzs++sI551KE+YU+55xLLd594ZpN9z7lXHLbErr0qMRi8O+HuvKv+7onO6ztlcfodeUiVGlQZWw+vID13+259e1OT62my0NfsPTefYh1in6FOv1zFfmvrIUMWPODPmwZFh7kUBmj8L7l5M7bBBLrxvRk8+EFW+tq/856ut+yhBXXDaJ8UPtWPc3GyMgw/vz8xxSvyOaKs/ZIdjg7rUOnKn5+01IG7L0FM7jl4t0o25LBRdcvo11ujKpKccdl/Zj/fup+FuCjL5qdpIuIprqbaWZntkD99wPPmNnk5q67qaoqxYRr+rBwdnvyOlRxx/MfM/ONjixZkJvs0LbJFiuvHIjlZkKl0euKTygd1pHyPduTWVRO7uyNVHbL3rb5si10+O96lt8yhKy1lfT4/WKW37YnZIiCJ1YTK8hi+W17QczI2Fi1dT+VVtHxuSLKhuQl4ywb5ZT/LWLpglza51c1vHEKO++az5n+Wkf+cM4AsrJj5OQZv73nUx66pSfTX+3EoceVMO53y/nVqYOTHWqdzNp2Uk5Wx8z5wNfiE3KYlb/NW7Mqm4Wzo1ZI6aZMli7MpVvviiRHVYMUJWRAVVFrufq6SpeJK1h7Zq/tHnKT924Jm75UANkZVPZoR2WvdrRbuBmA/FfXsv6U8IizDG1tWQN0nrSSkm92x7JTu3+wW+9yRowq4bmHC5MdSpO0z69i/8M38Xw4j8qKDDaVZGIGHTpGXzYdOlWxZmV2fdWkhOaY5D5VtXoilHQ3sAfwlKTdgUnAAKBI0k+Bu4Hdw+Y/M7O3JF0VyvYIP/9kZreH+sYCvyQavjjLzP4n7Hu0pIuBXsCvUrHV3LNfOYP2K+WjmSn4p2LM6P3rhWR9Uc6G0YWUD2lP3vQSqgqzqRiwfcs2c00F5UO2nUNVYTZZayqp2BT9R+886Qty5m2ismc71vywD7HO2WQvLiWzqILSQzrR6ek6nyGZEn589XLu/UNv2ufHkh1Kk/TqX8764kx+cetS9ti3lAWz2jP+8j7cfUVf/u+RRfzoihVIxs+/OSTZoTaoLfcpt3oTxcx+TPT8qmOBW4FDgJPN7AzgNuBWMzsU+A7bP/Z7b2A0MAK4UlK2pH2B3wLHmdmBwE/jtu8NHAmcRPQkgVpJOkfSdEnTKyhrrtNsUG77Ki6/91PuvqIPmzdmttpxE5YhVvxxCMvu3pucT0rJ/qyUgidWse70njtsqjr+g6jKyCquoGyvDnxxwxDK9mxPlwe/gJhROHEFa8f2buGTaLrDji9hXVHW1r9u0llmpjF4/1KeeaArF3x1L7ZszuD0C1dx0lnF3HNlH74/fCj3XNWXi29ZmuxQ62WIWCwjoSUdpULUT5lZaVg/HrhD0vtEz7bqJKn60d/PmlmZmRURPciwJ3AcMDmUYWbxDyT8l5nFzGxe2LZWZjbBzIab2fBscpr3zOqQmWVcfu+nvPJEF956rnOrHHNnWYdMtgztQPt3S8haVU6fSxbQ94KPyCyuoPevF5KxroLKrtlkFm/rgslcU0FlYRaxjpnEcsTmEZ0A2Hx4Ae0Wl6ItMbKXbqHX1Yvoe8FH5CzYTPcbP6PdJ5uTdZp1GnroJg7/agkTp87jsvGfceCRG/nVnz9Ldlg7pWhFNqtXZDP/vQ4AvPlMAYP3L+Urp63hzX9HF1/feLqAPYel3udQkyW4pKNU6MfdFLeeAYyMS9IARM8i3K4ZW0UUu6j73z5++xTqXDIuvnkpSxfk8sSEFBt1EWSUVGKZwjpkovIYubM3UnJyd5bdO3TrNn0v+IgV1w0m1imL0uGd6Hb7UkpO6kbW2kqyVpRRPrg9SJQe0onceZvYsl8+uXM2UdEvB2ufybL7ttXV86pFrP2fXik5+uJv1/Xmb9dFLfoDRm7k1B+v4saf9E9yVDtn7epsipa3o9+gLSz7JJdhR21kyYJcevcv54CRm5j1dj7DjtzI8sWt0zjZaW38Ql8qJOV4LwAXAn8EkDTMzN6vZ/uXgX9KutXMiiUV1mgtp5x9R2zi+NPWsmheLne9OB+I/uO/+0qnJEe2TebaCrrduQxigBmbRxZQekjd8VXslsumkQX0uXhBNCRuXF/IiP7TrD2zF93uWEqX+1dQ1SmT4vP71VmPa3l3/q4vv75jCVnZxhdL2nHzz3fj7SmdOO+a5WRmGuVlGfzpkjT4jNK1GZyAVEvKFwF3SppFFNsbwI/r2jg8XfZa4HVJVcB7wNmtEejOmjstn9F9Dkx2GPWq6J/Hihvrv9jz+Z17b/e65Ns9KPl2jx22q+rejpVXD6q3rpVXpce431lv5zPr7fxkh9Eki+bm8ZMT99yubO60fC48Yc869khNu2RLWdKfqef7yMwu2tmDmtmAsHpVjfIi4PRatq+53X5x6xOBiTXeP7vG6/T+n+Sc28qAWGwXTMrA9FaLwjnnEmXArthSDi3QrSR1MLNNdW3vnHOtZZcepyxppKR5wIfh9YGS7mrxyJxzri5teExcIuOU/0R000YxgJl9ABzdgjE551w9hFliSzpKaPSFmS0NY4WrpfesLM659JamreBEJJKUl0r6EmCS2hENW/uwZcNyzrk6GFgbHn2RSPfFj4ELgL7A58Cw8No555JECS7pp8GWchg73OxzHjvn3E5rw90XiYy+2EPS05JWS1ol6UlJ6XELlnOubdrFR188DDxGNBVmH+Bx4JGWDMo55+pUffNIIksDJOVKmibpA0lzJV0dygslvShpQfjZJW6fyyQtlDRf0ui48kMkzQ7v3a4wOkJSjqRJoXyqpAH1xZRIUpaZPWhmlWF5iLT9DnLOtQXRI6EaXhJQxrb52IcBJ0g6HLgUeNnMhhBNfHYpgKShwBhgX+AE4C5J1ROijwfOAYaE5YRQPg5Ya2aDieaQv6G+gOpMyuGbohB4VdKlkgZI6i/pV8CzCZ2uc861hJgSWxpgkY3hZXZYDDiZbXPqTAROCesnA4+Gud0XAwuBEZJ6A53M7G0zM+CBGvtU1zUZGFXdiq5NfRf6ZoTgqnc+N/5cgN/Xs69zzrWYup52U4tukuLn8ZlgZhO2qytq6c4ABgN3mtlUST3NbAWAma2QVD0FYl/gnbjdl4WyirBes7x6n6WhrkpJ64GuQK3PQatv7ouB9Z2pc84lReMu4hWZ2fB6qzOrAoZJ6kw0P/t+9WxeWwvX6imvb59aJXRHXwhyKJC7tUazBxLZ1znnmldiF/Eay8zWSXqNqC94paTeoZXcm+gRdBC1gHeL260f0TNHl4X1muXx+yyTlAUUAHU+jCORIXFXAn8Oy7HAjcA3G9rPOedaTDMNiZPUPbSQkZRH9JzQj4ieEXpW2Ows4Mmw/hQwJoyoGEh0QW9a6OrYIOnw0F88tsY+1XWdCrwS+p1rlUhL+VTgQOA9M/uBpJ5s/5Rp55xrXbFmq6k3MDH0K2cAj5nZM5LeBh6TNA5YApwGW5929BgwD6gELgjdHwDnAfcDecBzYQG4D3hQ0kKiFvKY+gJKJCmXmllMUqWkTkTNeL95xDmXHM04yb2ZzQIOqqW8GBhVxz7XAtfWUj4d2KE/2sy2EJJ6IhJJytND8/4vRFcoNwLTEj2Ac841t0aMvkg7icx9cX5YvVvS80Rj8Wa1bFjOOVePXTEpSzq4vvfMbGbLhOScc7uu+lrKN9fzngHHNXMsroX0/+7sZIfQYvJe75nsEFpE6ZdXJjuElLZLdl+Y2bGtGYhzziXESOgW6nSV0M0jzjmXUnbFlrJzzqWqXbL7wjnnUlYbTsqJ3GYtSd+XdEV4vbukES0fmnPO1WEXf/LIXcBI4Hvh9QbgzhaLyDnn6iFLfElHiXRfHGZmB0t6D8DM1kpq18JxOedc3Xbx0RcVYbIOg2hWJZpzOhDnnGukdG0FJyKR7ovbgX8CPSRdC7wJ/F+LRuWcc/Vpw33Kicx98XdJM4hmTBJwipl92OKROedcbdK4vzgRDSZlSbsDm4Gn48vMbElLBuacc3XalZMy0ZOrq59BlQsMBOYTPWLbOedandrwVa1Eui/2j38dZo87t47NnXPONUGj7+gzs5mSDm2JYJxzLiG7cveFpIvjXmYABwOrWywi55yrz65+oQ/oGLdeSdTH/I+WCcc55xKwqyblcNNIvpld0krxOOdcw3bFpCwpy8wq63sslHPOtTax646+mEbUf/y+pKeAx4FN1W+a2RMtHJtzzu3I+5QpBIqJnslXPV7ZAE/Kzrnk2EWTco8w8mIO25JxtTb8T+KcS3ltOAPVl5QzgXy2T8bV2vA/iXMu1e2q3RcrzOyaVoukDbv4liUcdvwG1hVlce5xewHQsXMlv7n7M3r2K2flsnZce25/Nq5P76dzfetHqznxjGLMxOKPcrn557tRUZbIRIStx6qMsnPWoO4Z5FzfpUl1VT5fSuUD0WWWrLEdyDohb7v3y/9UQtXzW8h7vkeTjtOSThm3mhPPXINkPPf3rvzz3u7JDikxbTgp1/c/pu3OIt3KXphUyG/PHLhd2XcvXMV7b+bzwyP34b038zn9wlVJiq55dO1VwSnjirjwxD0597i9yMwwjjl5XbLD2kHl5M1k9G/cl1/ZT9cQW1G1XZmVxKi8fxM5dxeSc08hlfdvwjZsGxIQ+6gCNqZ25ui/VyknnrmGi74+hB8fvxeHfaWEPgPLkh1WwywafZHIko7qS8qjWi2KGiSld5OxhjlT89mwdvtTGjm6hJceKwTgpccKGXlCSTJCa1aZWUZOboyMTCMnL0bxyuxkh7QdW1VF7J1yMk/a1qKNfV5J2SVr2fKjYsouXEPss8qE6qqaVkbG8HaoUwbqmEHG8HZUTY0SmlUZFeM3kH1efoucR3PZfUgZH85sT1lpBrEqMevtfI44cX2yw0rMrjifspmtaWrlki4HzgSWAkXADKIJ8+8EuhNNCfojM/tI0v3AGuAgYKakrkApsDfQH/gBcBbR8wKnmtnZ4RjjgUOBPGCymV0Zyj8FJgLfALKB08zso6aeU3Pp0q2CNauipLVmVTaduyaWDFJV8RfZTB7fnQff/ZCyLWLm6x2Z+XrHhndsReV3bCD7x/nY5m3/Wytu2kD2LzqS0S+L2LwKKm4tIedPhQ3WZUUx1GNbm0bdM7CiqGlW9c9SMo/IQV0zm/8kmtGnH+Vy9q9X0LFLJeVbMjj0uBIWzMpreMcU0Jb7lFusw0/ScOA7REn228Dw8NYE4CdmdgjwS6IHs1bbEzjezH4RXnchGor3c6L5nG8lmjJ0f0nDwja/NbPhwAHAlyUdEFdfkZkdDIwPx6otznMkTZc0vYI0+NMtReUXVDJydAlnHbYPZxy0L7ntYxz37bXJDmurqv+Woc4ZZOy1rfVum2PE5pRTfuV6towrpvzmEqw4SqyV/y5ly7hitowrJja/kvJfr2XLuGLKfrsu7FzLQQRWVEXVa1vI/Hb7lj+pJlq6MJfH7urBdY8u4tq/L2LxvDyqKtOk17KZWsqSdpP0qqQPJc2V9NNQXijpRUkLws8ucftcJmmhpPmSRseVHyJpdnjvdkkK5TmSJoXyqZIG1BdTS3YTHAk8aWalIbCnieZj/hLweIgXICdun8fNLL7z7mkzM0mzgZVmNjvUNRcYALwPfFfSOeFcegNDgVlh/+qx1DOIvhh2YGYTiL4o6KTCVvv+XVuUTWGPqLVc2KOCdcXp3WNz0FEb+WJpO9avic7jrX8XMHT4Jl55omkX05pLbE45Vf8tY8vU1Vg5sClG+f+VQH4Gufd13WH7rK/lkfW1qNVY9tM1ZF9aQEbvbS1fdc8g9n7F1te2OkbGsGxiCyqJfV5F2ZlF0RtbjC1nFJH7cLcWPb+dNeWRrkx5JDr/H1y6gtUrUqvLqVbN2zVRCfwizH7ZEZgh6UXgbOBlM7te0qXApcCvJQ0FxhA1DvsAL0naM+St8cA5wDvAv4ETgOeAccBaMxssaQxwA3B6XQG15KXx2r5yM4B1ZjYsbtkn7v1NNbavbrrG4tarX2dJGkjUAh5lZgcQTZaUW8v+VbTsF1CjvfNCJ47/btRDdPx31/D2lE5JjqhpVn2ezT4HbyInLwYYw47cyJKFOQ3u11qyz+lI3uTu5E7qTrsrCsg4uB05f+hMRu9Mql7dAoCZEVtY0UBNkcwROcTeLcM2xLANMWLvlpE5IofMkTnk/TM6Tu6k7pCrlE3IAAVdo/Pt3recI762ntf+1Tm5ASVARN0XiSwNMbMVZjYzrG8APgT6AicTdX8Sfp4S1k8GHjWzMjNbDCwERkjqDXQys7fNzIAHauxTXddkYJTiWqU1tWSiehO4R9J14ThfB/4CLJZ0mpk9HgI7wMw+2MljdCJK5Osl9QROBF5reujN69K7PuOAkRspKKzkoenzePDmnky6owe/vfszThizhlWfR0Pi0tn89zrwn2c7c+eUj6mqFAvn5PHcQzu2QFNN9uUFVNxSQsWDm6DSyDwul4zBDbcW1SmDrLH5lJ0bfbFmnZWPOqXW8L9EXHHvZ3TsUklVhbjjN33TZlhmI/qUu0maHvd6QvjreMc6o26Fg4CpQE8zWwFR4pZUPa6xL1FLuNqyUFYR1muWV++zNNRVKWk90JXoOtsOWuwTMLN3w5wZHwCfAdOB9UQX/sZL+h3RBbhHwzY7c4wPJL0HzAUWAW81R+zN7frza0+4l54+qJUjaVkP3tSLB2/qlewwGpR5UDsyD2oHQEbvTHL+WH8XS85ttV/4y/p6Hllfr//CWCqPUQb4xbcGJzuEnZN4Ui4K15zqJSmfaErin5lZST0N2bpupqvvJrtG3YDX0l+LN5nZVZLaA28AN4cm/wk1N6weTVHbazP7FNivjve22y+ufEDc+nTgmJ05AedcCmrGqz+SsokS8t/jJlpbKal3aCX3BqpvJFgG7Ba3ez9geSjvV0t5/D7LwnDfAqKRZrVq6b+3Jkh6H5gJ/KO678Y553Zagv3JiXRxhC7U+4APzeyWuLeeIhqCS/j5ZFz5mDCiYiAwBJgWujo2SDo81Dm2xj7VdZ0KvBL6nWvVoi1lMzujJet3zu2imq+lfATwP8Ds0IAE+A1wPfCYpHHAEuA0ADObK+kxYB7RyI0L4kaMnQfcT3TPxHNhgSjpPyhpIVELeUx9AaVHr75zzsVprluozexN6p5Sota7ms3sWuDaWsqnE9fNGle+hZDUE+FJ2TmXdtryHX2elJ1z6SWN57VIhCdl51z68aTsnHOpofqOvrbKk7JzLu0o1nazsidl51x68T5l55xLLd594ZxzqcSTsnPOpQ5vKTvnXCrxpOyccynC0vdJ1YnwpOycSys+Ttk551JN3TNfpj1Pys65tOMtZeecSxV+84hzzqUWv9DnnHMpxJOyc86lCsMv9DmXqkq/vDLZIbSIpZd/KdkhtIjyv7zTLPX4hT7nnEslnpSdcy41+M0jzjmXSsx8knvnnEspbTcne1J2zqUf775wzrlUYYB3XzjnXAppuznZk7JzLv1494VzzqUQH33hnHOpwmeJc8651BHdPNJ2s3JGsgNwzrlGiyW4NEDSXyWtkjQnrqxQ0ouSFoSfXeLeu0zSQknzJY2OKz9E0uzw3u2SFMpzJE0K5VMlDWgoJk/Kzrm0I7OElgTcD5xQo+xS4GUzGwK8HF4jaSgwBtg37HOXpMywz3jgHGBIWKrrHAesNbPBwK3ADQ0F5EnZOZderBFLQ1WZvQGsqVF8MjAxrE8ETokrf9TMysxsMbAQGCGpN9DJzN42MwMeqLFPdV2TgVHVrei6eJ+ycy7NNGrui26Spse9nmBmExrYp6eZrQAwsxWSeoTyvkD83KPLQllFWK9ZXr3P0lBXpaT1QFegqK6De1J2zqWfxC/0FZnZ8GY6am0tXKunvL596uTdF8659GLR46ASWXbSytAlQfi5KpQvA3aL264fsDyU96ulfLt9JGUBBezYXbIdT8rOufRjltiyc54CzgrrZwFPxpWPCSMqBhJd0JsWujo2SDo89BePrbFPdV2nAq+Efuc6efeFcy79NNMwZUmPAMcQ9T0vA64ErgcekzQOWAKcBmBmcyU9BswDKoELzKwqVHUe0UiOPOC5sADcBzwoaSFRC3lMQzF5UnbOpR3Fmudx1mb2vTreGlXH9tcC19ZSPh3Yr5byLYSknihPys659GIkdGNIuvKk7JxLKyLhG0PSkiflVta9TzmX3LaELj0qsRj8+6Gu/Ou+7skOa6dcfMsSDjt+A+uKsjj3uL0A2GNoKT+5fhl5HWKsXNaOGy7Ync0bMxuoKXWly+fVq8NGrhv1Mt3ab8ZMPDZvKA/NPoCfHDqN4wYuxkwUl+bxm1eOY/XmDvTpWMIzYx7l03WdAfhgZU+ufuPLAGRnVPHbo/7DiD7LiZm4bdoIXlw0iNOHzuV7+80hZmJTRTZXvf5lPllbmJwT9qSceiT918y+VM/7G80svzVjSkRVpZhwTR8Wzm5PXocq7nj+Y2a+0ZElC3KTHVqjvTCpkKf+1o1Lblu6texnNy3lL9f0YfY7+Xx1TDGnnreKB/7YO4lRNk26fF6VJm7875f4sKg77bPLmXzqZN5e1o+/vj+MP787AoDv7z+L84dP35p8l5Z04tuPf3eHus49ZAZrSvP42iNnIIyC3C0APLNgCJPm7QvAsQMW86sv/Zdznz2plc6whjaclNN2SFx9CTmVrVmVzcLZ7QEo3ZTJ0oW5dOtdkeSods6cqflsWLv993q/QWXMfqcDAO+90ZEjv74+GaE1m3T5vIo2d+DDoqgFv7miHYvWdqFHh01sqmi3dZu8rMqEBi18a++P+MvMgwEwxLoteQA71JU01X3KzTAhUSpqsZaypA7AY0QDqTOB3xNNxjEJODZsdoaZLZT0DeB3QDugGDjTzFZKugrYHdgj/PyTmd0e6t9oZvlhcPckoFM4n/PM7D9hm2uBk4BS4GQzW9lS57szevYrZ9B+pXw0s32yQ2k2n83PZeToEt6eUsBRJ62ne5/US2A7K10+rz4dS9inWxGzVvYE4KcjpvLNveazsbwdZz958tbt+nbcwD9OfZyNFdncPm0EM1b0oWO7MgB+MmIaI/osZ2lJJ/7wn6MoLo3O+Xv7zuGsAz8gO7OKHz71zdY/uaC5Rl+kopZsKZ8ALDezA81sP+D5UF5iZiOAO4A/hbI3gcPN7CDgUeBXcfXsDYwGRgBXSsqucZwzgClmNgw4EHg/lHcA3jGzA4E3gB8136k1XW77Ki6/91PuvqJPWve51nTLxbvxjbOLuOP5j8nLr6KyvN65V9JGunxe7bMquG30FK5764itLdvbph3GqAfH8szHe3Lm/rMBWL2pA6Me/B++M/k0bnjrCG48/iU6ZJeTmRGjd/4m3vuiF6dOPo33v+jFJSPf3lr/I3P344SHz+SWdw7n3ENmJOUcIcEbR9K0i6Mlk/Js4HhJN0g6ysyq/459JO7nyLDeD5giaTZwCdHUeNWeDbMyFRHd7tizxnHeBX4QWtX7m9mGUF4OPBPWZwADagtS0jmSpkuaXkHZzpxno2VmGZff+ymvPNGFt57r3CrHbC1LF+bym+8N4sIT9uS1f3VhxWftGt4pxaXL55WVUcWfRk/hmY/35KXFe+zw/rMLhvCVPRYBUBHLZH1Z1C8+r6g7S9cXMKDzOtZtyWVzRRYvLYr2n/LJIIZ2X71DXf9eMIRRAz5tuZOpj+FJeWeY2cfAIUTJ+TpJV1S/Fb9Z+Pln4A4z2x84F4i/ihKfKauo0eUSpt47Gvic6M6ZseGtirjbGXfYL27/CWY23MyGZ5PTmFPcScbFNy9l6YJcnpiQelfxm6qga9RdIRln/HQlzzzYNckRNVW6fF7G7495jUXrOjNx1oFbS/sXrNu6fuyAT1m0NpqvvUtuKRlhcoh+HUvoX7CeZSWdAPHapwMY0fdzAA7vt4xPwj7xdX25/2d8tr6gZU+pPt6n3HiS+gBrzOwhSRuBs8NbpxPdxng6UP13UQFRUoVt94knepz+wOdm9pfQj30w0XymKWnfEZs4/rS1LJqXy10vzgfgb9f15t1XOiU5ssa79K7POGDkRgoKK3lo+jwevLknee1jfOPsaFbCt54r4IVHkzRkqpmky+d1cK8vOHmvj5lfXMgTpz0GwJ+mHsa39/mQgZ3XETOxfENHrn7jaACG91nOTw59l8pYBjETV79x9NaW8y3vHM71o17m0iPeYm1pHr99NboEdMZ+cxjZbxmVsQzWl+Xwm1eOS87J0rYfB9WSQ+L2B/4oKUY03+h5RJM850iaStRKr77F8SrgcUmfE81XOrARxzkGuERSBbCRaDKQlDV3Wj6j+xzY8IZp4Prz+9danorjeHdWunxeM7/ozdDx5+1Q/saS2j+jFxcN4sVFg2p9b/nGjox98pQdyq9768gmxdisPCk3nplNAabEl4UJ9+80s6trbPsk22ZVii+/qsbr/eLW88PPiWyb2Z+a74f1yURfCM65dGcGVWnaN5GAtL15xDm3C/OWcvMwswGteTznXBvlSdk551KEAYk/oy/teFJ2zqUZA/M+ZeecSw2GX+hzzrmU4n3KzjmXQjwpO+dcqkjfeS0S4UnZOZdeDGjDU3d6UnbOpR9vKTvnXKrw26ydcy51GJiPU3bOuRTid/Q551wK8T5l55xLEWY++sI551KKt5Sdcy5VGFZVlewgWownZedcevGpO51zLsX4kDjnnEsNBpi3lJ1zLkWYT3LvnHMppS1f6JO14aEljSVpNfBZKx2uG1DUSsdqTW31vKDtnltrnld/M+velAokPU8UcyKKzOyEphyvtXlSThJJ081seLLjaG5t9byg7Z5bWz2vdJWR7ACcc85t40nZOedSiCfl5JmQ7ABaSFs9L2i759ZWzysteZ+yc86lEG8pO+dcCvGk7JxzKcSTcpJIGiBpTlgfLun2erY9RtIzrRcdSLpI0oeS/t5C9d8v6dSWqNs1naT/NvD+xtaKZVfjd/SlADObDkxPdhw1nA+caGaLqwskZZlZZRJjalW72vnGM7MvJTuGXZW3lHeCpO9LmibpfUn3SMqUtFHStZI+kPSOpJ5h20Hh9buSrqmthRHfEpb05VDv+5Lek9QxbJYvabKkjyT9XZJa8PzuBvYAnpK0XtIESS8AD0jqLukf4XzelXRE2OcqSX+V9JqkRZIuiqtvrKRZ4d/mwbhDHS3pv2H7Fms1S7o8/Lu9KOkRSb8Mn8vzkmZI+o+kvcO290u6RdKrwA3h9XhJr4Y4vxzO80NJ98cdY7yk6ZLmSro6rvxTSVdLmilpdvVxmuGcOkh6NvybzpF0ejjWDeF3c5qkwWHbb0iaGn6fXor73azvM9sYfvaW9Eb4fZwj6ai4bXb4fXfNwMx8acQC7AM8DWSH13cBY4kmr/pGKLsR+F1Yfwb4Xlj/MbAxrA8A5oT1Y4BnwvrTwBFhPZ/or5ljgPVAP6Iv0reBI1v4PD8lupX1KmAGkBfKH64+NrA78GFYvwr4L5AT9isGsoF9gflAt7BdYfh5P/B4OJ+hwMIWOo/hwPtAHtARWAD8EngZGBK2OQx4JS6uZ4DMuNePAgJOBkqA/UPcM4BhNc4rE3gNOCDu3/EnYf184N5mOq/vAH+Je10QjvXb8Hps3O9UF7aNtPpf4Ob6PrPwXvXv6S/i6swEOob1Wn/ffWn64t0XjTcKOAR4NzRW84BVQDnRf2aI/rN+JayPBE4J6w8DNzVQ/1vALaEv9wkzWxaOM83MlgFIep8oqb/Z5LNJzFNmVhrWjweGxjXUO8W15p81szKgTNIqoCdwHDDZzIoAzGxNXL3/suhZ8fNasKV1JPBkdfySngZygS8Bj8edR07cPo+bWfyMN0+bmUmaDaw0s9mhrrlEn8P7wHclnUP0Jdqb6ItmVtj/ifBzBvDtZjqv2cBNkm4gSr7/CefySHj/EeDWsN4PmCSpN9AOWBxXT22f2bK4998F/iopm+jzej+U1/X77prIuy8aT8BEMxsWlr3M7CqgwkKzAahiJ/vrzex6otZMHvBO3J+7ZXGb7XT9O2lT3HoGMDLu/Pua2YbwXm0xiqhVVZv47VuqO6a2ejOAdXHnMMzM9ol7f1ON7avjjLF9zDEgS9JAotb3KDM7AHiWKPHX3L/ZPjcz+5iocTAbuE7SFdVvxW8Wfv4ZuMPM9gfOrSO2WuMzszeAo4HPgQcljQ1vNcvvu9uRJ+XGexk4VVIPAEmFkvrXs/07RH9qAoxpqHJJg8xstpndQHTxr1n6IJvRC8CF1S8kDWtg+5eJWpFdw/aFLRdard4EviEpV1I+8HVgM7BY0mkhJkk6sAnH6ESUyNeHFv+JTQ26IZL6AJvN7CGiv74ODm+dHvfz7bBeQJRUAc5q5HH6A6vM7C/AfXHHcS3Ev90ayczmSfod8IKkDKACuKCeXX4GPCTpF0QtqPUNHOJnko4lan3MA54j6gJJFRcBd0qaRfT78wZRX3mtzGyupGuB1yVVAe8BZ7dGoOH470p6CviAaFrW6USfwZnA+PBZZhP1G3+wk8f4QNJ7wFxgEVEXVEvbH/ijpBjR7+B5wGQgR9JUogbX98K2VxF11XxO1EgY2IjjHANcIqkC2EjUV+1akN9m3cIktQdKQ5/kGKKLficnO65diaR8M9sYPos3gHPMbGay42pukj4Fhlf337v05C3llncIcIeiqzDrgB8mN5xd0gRJQ4n6Uie2xYTs2g5vKTvnXArxC33OOZdCPCk751wK8aTsnHMpxJOyS5ikqrg5EB4Poxl2tq6ts8RJujdciKtr22MkNXqCnDAXxA5PPa6rvMY2jZoFLcwj8cvGxuhcTZ6UXWOUhrvf9iO6zXa78cmSMnemUjP7XzObV88mxxDdFu1cm+dJ2e2s/wCDQyv2VUkPA7MVzZj3R0UzyM2SdC5svWvuDknzJD0L9KiuKMxSNjysn6BoRrUPJL0saQBR8v95aKUfpbpnqusq6QVFs6HdQwK3bkv6l6KZ4uaGuSvi37s5xPKypO6hrNbZ5ZxrLj5O2TWapCyiW4mfD0UjgP3MbHFIbOvN7FBJOcBbiqb9PAjYi+hOtJ5Edyv+tUa93YG/AEeHugrNbI2iqUQ3mtlNYbuHgVvN7E1JuwNTiGbvuxJ408yukfR1YLskW4cfhmPkEU0y9Q8zKwY6ADPN7BdhXokriW4vnwD82MwWSDqMaJbA43bin9G5WnlSdo2Rp2iGOohayvcRdStMs22T4X8VOEDb5kcuAIYQTWrzSJh9bbmkV2qp/3Dgjeq6aswoF6+umeqOJszCZmbPSlqbwDldJOlbYX23EGsx0WRDk0L5Q8ATYe6M+maXc67JPCm7xig1s2HxBSE5xc+qJqL5g6fU2O5r1D1bXPy+idzNVD1TXWl8YYgl4buhJB1DlOBHmtlmSa+x/Qxq8Yy42eUSPYZzjeV9yq65TQHOUzT/LpL2lNSBaM6JMaHPuTdwbC37vg18WdFUmPEzym0gmqC+Wl0z1b1BNNEQkk4kmty9PgXA2pCQ9yZqqVfLAKpb+2cQdYuU0Lyzyzm3A0/KrrndS9RfPFPRg2HvIfqL7J9ET/2YDYwHXq+5o5mtJuoHfkLSB2zrPnga+Fb1hT6imeqGhwuJ89g2CuRqokdMzSTqRlnSQKzPE82HPAv4PdEMatU2AftKmkHUZ3xNKD8TGBfim0v0NBLnmo3PfeGccynEW8rOOZdCPCk751wK8aTsnHMpxJOyc86lEE/KzjmXQjwpO+dcCvGk7JxzKeT/AS57CUZrt2D0AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "ConfusionMatrixDisplay((confusion_matrix(clf.predict(X_test), y_test)), display_labels=le.inverse_transform(clf.classes_)).plot()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "180f89db",
   "metadata": {},
   "source": [
    "The accuracy of the model is quite good and most of the predictions lie in the diagonal of the confusion matrix. Let us try o predict a few sentences in 4 languages and check how the model predict sentences with foregn borrowed words."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "id": "ebf1a207",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "I will have pasta and pizza for dinner                  -    English with confidence 99.91%\n",
      "Te hago un forward con toda la info                     -    Spanish with confidence 99.95%\n",
      "Puis-je accéder à Internet avec Skype Zones             -     French with confidence 99.94%\n",
      "Ich muß nach England telefonieren                       -     German with confidence 99.91%\n",
      "Voy del aeropuerto al hospital                          -    Spanish with confidence 99.95%\n",
      "He has already called the doctor                        -    English with confidence 99.91%\n",
      "Juan no viene al restaurante hoy                        -    Spanish with confidence 99.95%\n",
      "Ci vediamo nel weekend                                  -     French with confidence 83.17%\n",
      "Facciamo una call conference in remote meeting su skype -    English with confidence 99.90%\n",
      "Espresso or cappuccino? I can't choose                  -    English with confidence 99.91%\n"
     ]
    }
   ],
   "source": [
    "phrases = [\"I will have pasta and pizza for dinner\", \n",
    "           \"Te hago un forward con toda la info\",\n",
    "           \"Puis-je accéder à Internet avec Skype Zones\",\n",
    "           \"Ich muß nach England telefonieren\",\n",
    "           \"Voy del aeropuerto al hospital\",\n",
    "           \"He has already called the doctor\", \n",
    "           \"Juan no viene al restaurante hoy\",\n",
    "           \"Ci vediamo nel weekend\", \n",
    "           \"Facciamo una call conference in remote meeting su skype\",\n",
    "           \"Espresso or cappuccino? I can't choose\"]\n",
    "\n",
    "for phrase in phrases:\n",
    "    pred_array = clf.predict(cv.transform([phrase]))\n",
    "    pred = le.inverse_transform(pred_array)\n",
    "    pred_proba = clf.predict_proba(cv.transform([phrase]))\n",
    "    ix = pred_proba.argmax(1)[0]\n",
    "    print(\"{:55} - {:>10} with confidence {:.2%}\".format(phrase, pred[0].title(), pred_proba[0,ix]))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f6122f27",
   "metadata": {},
   "source": [
    "The model is doing good on sentences with foreign borrowed words.\n",
    "\n",
    "Now I will use a char bigram CountVectorized to preprocess the text, i.e. a word like \"apple\" will be transformed into following char bigrams \"ap\", \"pp\", \"pl\" and \"le\". My guess is that this method will be more accurate as the model learns to spot two-char clusters, instead of the whole word. In other words, the model is taught by internal structure (phonotactics) of the language rather than by the words present in the training set, which implies that it will have higher accuracy even with small training sets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 251,
   "id": "e9e8d50a",
   "metadata": {},
   "outputs": [],
   "source": [
    "cv_char = CountVectorizer(analyzer = 'char_wb', ngram_range=(2,4))\n",
    "X_char = cv_char.fit_transform(X_raw)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 252,
   "id": "f648179c",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_char, X_test_char, y_train_char, y_test_char = train_test_split(X_char, y, test_size = 0.2, random_state = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 253,
   "id": "aab1f81f",
   "metadata": {},
   "outputs": [],
   "source": [
    "# another_clf = MultinomialNB()\n",
    "another_clf = SGDClassifier(loss = \"log\", alpha = 1e-05, random_state=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 254,
   "id": "19ee58e4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "SGDClassifier(alpha=1e-05, loss='log', random_state=1)"
      ]
     },
     "execution_count": 254,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "another_clf.fit(X_train_char, y_train_char)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 255,
   "id": "d6778d97",
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred_char = another_clf.predict(X_test_char)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 256,
   "id": "1ea75a48",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9994565822295842"
      ]
     },
     "execution_count": 256,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "accuracy_score(y_test_char, y_pred_char)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 268,
   "id": "c82e4961",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<sklearn.metrics._plot.confusion_matrix.ConfusionMatrixDisplay at 0x26348e9cc70>"
      ]
     },
     "execution_count": 268,
     "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": [
    "ConfusionMatrixDisplay((confusion_matrix(y_test, y_pred_char)), display_labels=le.inverse_transform(another_clf.classes_)).plot()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 258,
   "id": "c515e38d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      1.00      1.00     55258\n",
      "           1       1.00      1.00      1.00     34104\n",
      "           2       1.00      1.00      1.00     39727\n",
      "           3       1.00      1.00      1.00     23648\n",
      "\n",
      "    accuracy                           1.00    152737\n",
      "   macro avg       1.00      1.00      1.00    152737\n",
      "weighted avg       1.00      1.00      1.00    152737\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_test_char, y_pred_char))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "09a5d865",
   "metadata": {},
   "source": [
    "Let us use again the sentences above to predict tha language"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "id": "403ebbdb",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "I will have pasta and pizza for dinner                  -    English with confidence 100.00%\n",
      "Te hago un forward con toda la info                     -    Spanish with confidence 99.99%\n",
      "Puis-je accéder à Internet avec Skype Zones             -     French with confidence 99.97%\n",
      "Ich muß nach England telefonieren                       -     German with confidence 99.99%\n",
      "Voy del aeropuerto al hospital                          -    Spanish with confidence 99.96%\n",
      "He has already called the doctor                        -    English with confidence 100.00%\n",
      "Juan no viene al restaurante hoy                        -    Spanish with confidence 99.99%\n",
      "Ci vediamo nel weekend                                  -    English with confidence 74.09%\n",
      "Facciamo una call conference in remote meeting su skype -    English with confidence 71.14%\n",
      "Espresso or cappuccino? I can't choose                  -    English with confidence 99.23%\n"
     ]
    }
   ],
   "source": [
    "phrases = [\"I will have pasta and pizza for dinner\", \n",
    "           \"Te hago un forward con toda la info\",\n",
    "           \"Puis-je accéder à Internet avec Skype Zones\",\n",
    "           \"Ich muß nach England telefonieren\",\n",
    "           \"Voy del aeropuerto al hospital\",\n",
    "           \"He has already called the doctor\", \n",
    "           \"Juan no viene al restaurante hoy\",\n",
    "           \"Ci vediamo nel weekend\",\n",
    "           \"Facciamo una call conference in remote meeting su skype\",\n",
    "           \"Espresso or cappuccino? I can't choose\"]\n",
    "\n",
    "for phrase in phrases:\n",
    "    pred_array = another_clf.predict(cv_char.transform([phrase]))\n",
    "    pred = le.inverse_transform(pred_array)\n",
    "    pred_proba = another_clf.predict_proba(cv_char.transform([phrase]))\n",
    "    ix = pred_proba.argmax(1)[0]\n",
    "    print(\"{:55} - {:>10} with confidence {:.2%}\".format(phrase, pred[0].title(), pred_proba[0,ix]))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "adad2863",
   "metadata": {},
   "source": [
    "This model seems to be more prone to errors when a sentence has forign words present. "
   ]
  }
 ],
 "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": "3.8.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}