notebooks/P2C3_regression_logistique_classification.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "6ea41722-9d8f-4e53-b08b-a03224bbb739",
   "metadata": {},
   "source": [
    "# P2C3: Classifiez avec la Régression Logistique\n",
    "\n",
    "\n",
    "[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/OpenClassrooms-Student-Center/8063076-Initiez-vous-au-Machine-Learning/blob/master/notebooks/P2C3_classification_regression_logistique.ipynb)\n",
    "\n",
    "Prenons un exemple avec le modèle de sklearn LogisticRegression\n",
    "\n",
    "Considérons le dataset du cancer du sein disponible directement dans sklearn. \n",
    "\n",
    "Ce dataset a 569 échantillons, 30 prédicteurs et une variable cible binaire : la tumeur est maligne (aïe, 1) ou bénigne (ouf, 0)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "e6456118-0a98-4dbe-addc-326610f2c54b",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.datasets import load_breast_cancer\n",
    "X, y = load_breast_cancer(return_X_y=True)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "65bd0789-c4b1-4e02-b5b6-118207111bb0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(569, 30)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "5bea7731-5483-4260-9614-6111ca1d1ad0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1,\n",
       "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0,\n",
       "       0, 0, 1, 0, 1, 1, 1, 1, 1, 0, 0, 1, 0, 0, 1, 1, 1, 1, 0, 1, 0, 0,\n",
       "       1, 1, 1, 1, 0, 1, 0, 0, 1, 0, 1, 0, 0, 1, 1, 1, 0, 0, 1, 0, 0, 0,\n",
       "       1, 1, 1, 0, 1, 1, 0, 0, 1, 1, 1, 0, 0, 1, 1, 1, 1, 0, 1, 1, 0, 1,\n",
       "       1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 1, 0, 0, 1, 1, 1, 0, 0, 1, 0, 1, 0,\n",
       "       0, 1, 0, 0, 1, 1, 0, 1, 1, 0, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1,\n",
       "       1, 1, 0, 1, 1, 1, 1, 0, 0, 1, 0, 1, 1, 0, 0, 1, 1, 0, 0, 1, 1, 1,\n",
       "       1, 0, 1, 1, 0, 0, 0, 1, 0, 1, 0, 1, 1, 1, 0, 1, 1, 0, 0, 1, 0, 0,\n",
       "       0, 0, 1, 0, 0, 0, 1, 0, 1, 0, 1, 1, 0, 1, 0, 0, 0, 0, 1, 1, 0, 0,\n",
       "       1, 1, 1, 0, 1, 1, 1, 1, 1, 0, 0, 1, 1, 0, 1, 1, 0, 0, 1, 0, 1, 1,\n",
       "       1, 1, 0, 1, 1, 1, 1, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
       "       0, 0, 1, 1, 1, 1, 1, 1, 0, 1, 0, 1, 1, 0, 1, 1, 0, 1, 0, 0, 1, 1,\n",
       "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 0, 1, 0, 1, 1, 1, 1, 1,\n",
       "       1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 0, 1, 0, 1, 1, 1, 1, 0, 0,\n",
       "       0, 1, 1, 1, 1, 0, 1, 0, 1, 0, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1, 0,\n",
       "       0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 1, 0, 0, 0, 1, 0, 0,\n",
       "       1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 0, 1, 1, 1, 0, 1, 1, 0, 0, 1, 1,\n",
       "       1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 0, 1, 1, 0,\n",
       "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 0, 0, 1, 0, 1, 1, 1, 1,\n",
       "       1, 0, 1, 1, 0, 1, 0, 1, 1, 0, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0,\n",
       "       1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1,\n",
       "       1, 1, 1, 0, 1, 0, 1, 1, 0, 1, 1, 1, 1, 1, 0, 0, 1, 0, 1, 0, 1, 1,\n",
       "       1, 1, 1, 0, 1, 1, 0, 1, 0, 1, 0, 0, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1,\n",
       "       1, 1, 1, 1, 1, 0, 1, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
       "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 1])"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "43ba9f2e-65dd-4ffb-a2fe-1da14b4ecce3",
   "metadata": {},
   "outputs": [],
   "source": [
    "# split train, test\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.20, random_state=42)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b16fee53-0e38-4172-8e07-6f0d5287f835",
   "metadata": {},
   "source": [
    "## Regression Logistique"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "f5f1149e-e4ea-4676-9b87-0c0ff1f68a06",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Library/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages/sklearn/linear_model/_logistic.py:458: ConvergenceWarning: lbfgs failed to converge (status=1):\n",
      "STOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n",
      "\n",
      "Increase the number of iterations (max_iter) or scale the data as shown in:\n",
      "    https://scikit-learn.org/stable/modules/preprocessing.html\n",
      "Please also refer to the documentation for alternative solver options:\n",
      "    https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n",
      "  n_iter_i = _check_optimize_result(\n"
     ]
    }
   ],
   "source": [
    "from sklearn.linear_model import LogisticRegression\n",
    "\n",
    "# initialisation et entraînement en une ligne\n",
    "clf = LogisticRegression(random_state=808).fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "3eaacee2-3ec2-45cd-8585-8ee32b50935a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Prédiction 0\n",
      "Probabilité 0.793373669887237\n"
     ]
    }
   ],
   "source": [
    "# Prediction d'un échantillon\n",
    "print(\"Prédiction\",clf.predict([X[8, :]])[0])\n",
    "\n",
    "print(\"Probabilité\",clf.predict_proba([X[8, :]])[0][0])\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b1b62013-61c3-4640-9399-664288349c78",
   "metadata": {},
   "source": [
    "On obtient 0, soit bénigne avec un probabilité de 79,3%\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "a266c599-44d4-40f0-badc-295ec63a1460",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Prédiction 1\n",
      "Probabilité 0.8734077641123704\n"
     ]
    }
   ],
   "source": [
    "# Prediction d'un autre échantillon\n",
    "print(\"Prédiction\",clf.predict([X[13, :]])[0])\n",
    "\n",
    "print(\"Probabilité\",clf.predict_proba([X[13, :]])[0][1])\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "93106f67-60e2-403d-ba3f-ae3de682a9b7",
   "metadata": {},
   "source": [
    "On obtient 1, soit maligne avec un probabilité de 87,3%\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "695dc80a-a958-49dd-b70b-194f85c7ef88",
   "metadata": {},
   "source": [
    "Un bon moyen d'analyser les performances d'un modèle de classification est de tracer l'histogramme des probabilités des prédictions.\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "b8006f63-b433-4c8a-9056-618e05edd207",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Axes: ylabel='Count'>"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjIAAAGdCAYAAAAIbpn/AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8pXeV/AAAACXBIWXMAAA9hAAAPYQGoP6dpAAAjUklEQVR4nO3de3BU9f3/8VdCrgqbEJBcJIGAQgDFC5cY0K+C0RSVwpCpqEjRIngJKGSqNlWMoAKlClQaoFAuOiNNpSNUFKESBKskCBEqCEQRbCKwq1GT5ZZNIOf3h8P+upIAWZOc/cTnY+bMuOecXd77KZpnT84mQZZlWQIAADBQsN0DAAAA+IuQAQAAxiJkAACAsQgZAABgLEIGAAAYi5ABAADGImQAAICxCBkAAGCsELsHaGq1tbU6fPiw2rRpo6CgILvHAQAAF8CyLB09elQJCQkKDq7/ukuLD5nDhw8rMTHR7jEAAIAfysrK1LFjx3qPt/iQadOmjaQfFsLhcNg8DQAAuBBut1uJiYner+P1afEhc+bbSQ6Hg5ABAMAw57sthJt9AQCAsQgZAABgLEIGAAAYi5ABAADGImQAAICxCBkAAGAsQgYAABiLkAEAAMYiZAAAgLEIGQAAYCxCBgAAGIuQAQAAxiJkAACAsVr8b78GAMBOpaWlKi8vt3uMJtO+fXslJSXZ9ucTMgAANJHS0lKlpPTQyZMn7B6lyURGXqR9+/baFjOEDAAATaS8vFwnT55Q6m9y5YjvbPc4jc595EttXTpV5eXlhAwAAC2VI76zYpK62z1Gi8TNvgAAwFiEDAAAMBYhAwAAjEXIAAAAYxEyAADAWIQMAAAwFiEDAACMRcgAAABjETIAAMBYhAwAADAWIQMAAIxFyAAAAGMRMgAAwFiEDAAAMBYhAwAAjEXIAAAAYxEyAADAWIQMAAAwFiEDAACMRcgAAABjETIAAMBYhAwAADAWIQMAAIxFyAAAAGMRMgAAwFi2h8yhQ4d07733ql27doqMjNSVV16p7du3e49blqVnnnlG8fHxioyMVHp6uj7//HMbJwYAAIHC1pD5/vvvNXDgQIWGhuqdd97Rnj179NJLL6lt27bec2bNmqWXX35ZCxcu1NatW3XxxRcrIyNDVVVVNk4OAAACQYidf/gf/vAHJSYmatmyZd59ycnJ3n+2LEtz587V008/rWHDhkmSXn31VcXGxmr16tW66667mn1mAAAQOGy9IvPmm2+qb9+++tWvfqUOHTrommuu0eLFi73HDx48KKfTqfT0dO++qKgopaamqrCwsM7X9Hg8crvdPhsAAGiZbA2ZAwcOaMGCBbr88su1fv16Pfzww3r00Uf1yiuvSJKcTqckKTY21ud5sbGx3mM/NmPGDEVFRXm3xMTEpn0TAADANraGTG1tra699lpNnz5d11xzjcaPH69x48Zp4cKFfr9mTk6OKisrvVtZWVkjTgwAAAKJrSETHx+vnj17+uzr0aOHSktLJUlxcXGSJJfL5XOOy+XyHvux8PBwORwOnw0AALRMtobMwIEDVVJS4rPvs88+U6dOnST9cONvXFycCgoKvMfdbre2bt2qtLS0Zp0VAAAEHls/tTR58mQNGDBA06dP15133qmPPvpIixYt0qJFiyRJQUFBmjRpkp5//nldfvnlSk5O1pQpU5SQkKDhw4fbOToAAAgAtoZMv379tGrVKuXk5GjatGlKTk7W3LlzNWrUKO85TzzxhI4fP67x48eroqJC119/vdatW6eIiAgbJwcAAIHA1pCRpDvuuEN33HFHvceDgoI0bdo0TZs2rRmnAgAAJrD9VxQAAAD4i5ABAADGImQAAICxCBkAAGAsQgYAABiLkAEAAMYiZAAAgLEIGQAAYCxCBgAAGIuQAQAAxiJkAACAsQgZAABgLEIGAAAYi5ABAADGImQAAICxCBkAAGAsQgYAABiLkAEAAMYiZAAAgLEIGQAAYCxCBgAAGIuQAQAAxiJkAACAsQgZAABgLEIGAAAYi5ABAADGImQAAICxCBkAAGAsQgYAABiLkAEAAMYiZAAAgLEIGQAAYCxCBgAAGIuQAQAAxiJkAACAsQgZAABgLEIGAAAYi5ABAADGImQAAICxCBkAAGAsQgYAABiLkAEAAMayNWSeffZZBQUF+WwpKSne41VVVcrKylK7du3UunVrZWZmyuVy2TgxAAAIJLZfkenVq5eOHDni3T744APvscmTJ2vNmjVauXKlNm/erMOHD2vEiBE2TgsAAAJJiO0DhIQoLi7urP2VlZVasmSJVqxYocGDB0uSli1bph49eqioqEjXXXddc48KAAACjO1XZD7//HMlJCSoS5cuGjVqlEpLSyVJxcXFqqmpUXp6uvfclJQUJSUlqbCw0K5xAQBAALH1ikxqaqqWL1+u7t2768iRI5o6dapuuOEG7d69W06nU2FhYYqOjvZ5TmxsrJxOZ72v6fF45PF4vI/dbndTjQ8AAGxma8gMGTLE+8+9e/dWamqqOnXqpNdff12RkZF+veaMGTM0derUxhoRAAAEMNu/tfS/oqOj1a1bN+3fv19xcXGqrq5WRUWFzzkul6vOe2rOyMnJUWVlpXcrKytr4qkBAIBdAipkjh07pi+++ELx8fHq06ePQkNDVVBQ4D1eUlKi0tJSpaWl1fsa4eHhcjgcPhsAAGiZbP3W0m9/+1sNHTpUnTp10uHDh5Wbm6tWrVrp7rvvVlRUlMaOHavs7GzFxMTI4XBo4sSJSktL4xNLAABAks0h89VXX+nuu+/Wt99+q0suuUTXX3+9ioqKdMkll0iS5syZo+DgYGVmZsrj8SgjI0Pz58+3c2QAABBAbA2Z/Pz8cx6PiIhQXl6e8vLymmkiAABgkoC6RwYAAKAhCBkAAGAsQgYAABiLkAEAAMYiZAAAgLEIGQAAYCxCBgAAGIuQAQAAxiJkAACAsQgZAABgLEIGAAAYi5ABAADGImQAAICxCBkAAGAsQgYAABiLkAEAAMYiZAAAgLEIGQAAYCxCBgAAGIuQAQAAxiJkAACAsQgZAABgLEIGAAAYi5ABAADGImQAAICxCBkAAGAsQgYAABiLkAEAAMYiZAAAgLEIGQAAYCxCBgAAGIuQAQAAxiJkAACAsQgZAABgLEIGAAAYi5ABAADGImQAAICxCBkAAGAsQgYAABiLkAEAAMYiZAAAgLEIGQAAYCxCBgAAGCtgQmbmzJkKCgrSpEmTvPuqqqqUlZWldu3aqXXr1srMzJTL5bJvSAAAEFACImS2bdumv/zlL+rdu7fP/smTJ2vNmjVauXKlNm/erMOHD2vEiBE2TQkAAAKN7SFz7NgxjRo1SosXL1bbtm29+ysrK7VkyRLNnj1bgwcPVp8+fbRs2TJt2bJFRUVFNk4MAAAChe0hk5WVpdtvv13p6ek++4uLi1VTU+OzPyUlRUlJSSosLKz39Twej9xut88GAABaphA7//D8/Hx9/PHH2rZt21nHnE6nwsLCFB0d7bM/NjZWTqez3tecMWOGpk6d2tijAgCAAGTbFZmysjI99thjeu211xQREdFor5uTk6PKykrvVlZW1mivDQAAAottIVNcXKyvv/5a1157rUJCQhQSEqLNmzfr5ZdfVkhIiGJjY1VdXa2Kigqf57lcLsXFxdX7uuHh4XI4HD4bAABomWz71tLNN9+sXbt2+ey7//77lZKSoieffFKJiYkKDQ1VQUGBMjMzJUklJSUqLS1VWlqaHSMDAIAAY1vItGnTRldccYXPvosvvljt2rXz7h87dqyys7MVExMjh8OhiRMnKi0tTdddd50dIwMAgABj682+5zNnzhwFBwcrMzNTHo9HGRkZmj9/vt1jAQCAABFQIbNp0yafxxEREcrLy1NeXp49AwEAgIBm+8+RAQAA8BchAwAAjEXIAAAAY/kVMl26dNG333571v6Kigp16dLlJw8FAABwIfwKmS+//FKnT58+a7/H49GhQ4d+8lAAAAAXokGfWnrzzTe9/7x+/XpFRUV5H58+fVoFBQXq3Llzow0HAABwLg0KmeHDh0uSgoKCNGbMGJ9joaGh6ty5s1566aVGGw4AAOBcGhQytbW1kqTk5GRt27ZN7du3b5KhAAAALoRfPxDv4MGDjT2HkUpLS1VeXm73GE2qffv2SkpKsnsMAADq5PdP9i0oKFBBQYG+/vpr75WaM5YuXfqTBwt0paWlSknpoZMnT9g9SpOKjLxI+/btJWYAAAHJr5CZOnWqpk2bpr59+yo+Pl5BQUGNPVfAKy8v18mTJ5T6m1w54jvbPU6TcB/5UluXTlV5eTkhAwAISH6FzMKFC7V8+XKNHj26secxjiO+s2KSuts9BgAAP0t+/RyZ6upqDRgwoLFnAQAAaBC/QuaBBx7QihUrGnsWAACABvHrW0tVVVVatGiRNmzYoN69eys0NNTn+OzZsxtlOAAAgHPxK2Q++eQTXX311ZKk3bt3+xz7Od74CwAA7OFXyLz33nuNPQcAAECD+XWPDAAAQCDw64rMoEGDzvktpI0bN/o9EAAAwIXyK2TO3B9zRk1NjXbu3Kndu3ef9cskAQAAmopfITNnzpw69z/77LM6duzYTxoIAADgQjXqPTL33nvvz+L3LAEAgMDQqCFTWFioiIiIxnxJAACAevn1raURI0b4PLYsS0eOHNH27ds1ZcqURhkMAADgfPwKmaioKJ/HwcHB6t69u6ZNm6Zbb721UQYDAAA4H79CZtmyZY09BwAAQIP5FTJnFBcXa+/evZKkXr166ZprrmmUoQAAAC6EXyHz9ddf66677tKmTZsUHR0tSaqoqNCgQYOUn5+vSy65pDFnBAAAqJNfn1qaOHGijh49qk8//VTfffedvvvuO+3evVtut1uPPvpoY88IAABQJ7+uyKxbt04bNmxQjx49vPt69uypvLw8bvYFAADNxq8rMrW1tQoNDT1rf2hoqGpra3/yUAAAABfCr5AZPHiwHnvsMR0+fNi779ChQ5o8ebJuvvnmRhsOAADgXPwKmT//+c9yu93q3Lmzunbtqq5duyo5OVlut1vz5s1r7BkBAADq5Nc9MomJifr444+1YcMG7du3T5LUo0cPpaenN+pwAAAA59KgKzIbN25Uz5495Xa7FRQUpFtuuUUTJ07UxIkT1a9fP/Xq1Uv//ve/m2pWAAAAHw0Kmblz52rcuHFyOBxnHYuKitKDDz6o2bNnN9pwAAAA59KgkPnPf/6jX/ziF/Uev/XWW1VcXPyThwIAALgQDQoZl8tV58euzwgJCdE333zzk4cCAAC4EA0KmUsvvVS7d++u9/gnn3yi+Pj4nzwUAADAhWhQyNx2222aMmWKqqqqzjp28uRJ5ebm6o477mi04QAAAM6lQR+/fvrpp/XGG2+oW7dumjBhgrp37y5J2rdvn/Ly8nT69Gk99dRTTTIoAADAjzUoZGJjY7VlyxY9/PDDysnJkWVZkqSgoCBlZGQoLy9PsbGxTTIoAADAjzX4J/t26tRJa9euVXl5ubZu3aqioiKVl5dr7dq1Sk5ObtBrLViwQL1795bD4ZDD4VBaWpreeecd7/GqqiplZWWpXbt2at26tTIzM+VyuRo6MgAAaKH8+hUFktS2bVv169dP/fv3V9u2bf16jY4dO2rmzJkqLi7W9u3bNXjwYA0bNkyffvqpJGny5Mlas2aNVq5cqc2bN+vw4cMaMWKEvyMDAIAWxq9fUdBYhg4d6vP4hRde0IIFC1RUVKSOHTtqyZIlWrFihQYPHixJWrZsmXr06KGioiJdd911dowMAAACiN9XZBrb6dOnlZ+fr+PHjystLU3FxcWqqanx+f1NKSkpSkpKUmFhYb2v4/F45Ha7fTYAANAy2R4yu3btUuvWrRUeHq6HHnpIq1atUs+ePeV0OhUWFqbo6Gif82NjY+V0Out9vRkzZigqKsq7JSYmNvE7AAAAdrE9ZLp3766dO3dq69atevjhhzVmzBjt2bPH79fLyclRZWWldysrK2vEaQEAQCCx9R4ZSQoLC9Nll10mSerTp4+2bdumP/3pTxo5cqSqq6tVUVHhc1XG5XIpLi6u3tcLDw9XeHh4U48NAAACgO1XZH6strZWHo9Hffr0UWhoqAoKCrzHSkpKVFpaqrS0NBsnBAAAgcLWKzI5OTkaMmSIkpKSdPToUa1YsUKbNm3S+vXrFRUVpbFjxyo7O1sxMTFyOByaOHGi0tLS+MQSAACQZHPIfP311/r1r3+tI0eOKCoqSr1799b69et1yy23SJLmzJmj4OBgZWZmyuPxKCMjQ/Pnz7dzZAAAEEBsDZklS5ac83hERITy8vKUl5fXTBMBAACTBNw9MgAAABeKkAEAAMYiZAAAgLEIGQAAYCxCBgAAGIuQAQAAxiJkAACAsQgZAABgLEIGAAAYi5ABAADGImQAAICxCBkAAGAsQgYAABiLkAEAAMYiZAAAgLEIGQAAYCxCBgAAGIuQAQAAxiJkAACAsQgZAABgLEIGAAAYi5ABAADGImQAAICxCBkAAGAsQgYAABiLkAEAAMYiZAAAgLEIGQAAYCxCBgAAGIuQAQAAxiJkAACAsQgZAABgLEIGAAAYi5ABAADGImQAAICxCBkAAGAsQgYAABiLkAEAAMYiZAAAgLEIGQAAYCxCBgAAGIuQAQAAxrI1ZGbMmKF+/fqpTZs26tChg4YPH66SkhKfc6qqqpSVlaV27dqpdevWyszMlMvlsmliAAAQSGwNmc2bNysrK0tFRUV69913VVNTo1tvvVXHjx/3njN58mStWbNGK1eu1ObNm3X48GGNGDHCxqkBAECgCLHzD1+3bp3P4+XLl6tDhw4qLi7W//3f/6myslJLlizRihUrNHjwYEnSsmXL1KNHDxUVFem6666zY2wAABAgAuoemcrKSklSTEyMJKm4uFg1NTVKT0/3npOSkqKkpCQVFhbaMiMAAAgctl6R+V+1tbWaNGmSBg4cqCuuuEKS5HQ6FRYWpujoaJ9zY2Nj5XQ663wdj8cjj8fjfex2u5tsZgAAYK+AuSKTlZWl3bt3Kz8//ye9zowZMxQVFeXdEhMTG2lCAAAQaAIiZCZMmKC33npL7733njp27OjdHxcXp+rqalVUVPic73K5FBcXV+dr5eTkqLKy0ruVlZU15egAAMBGtoaMZVmaMGGCVq1apY0bNyo5OdnneJ8+fRQaGqqCggLvvpKSEpWWliotLa3O1wwPD5fD4fDZAABAy2TrPTJZWVlasWKF/vnPf6pNmzbe+16ioqIUGRmpqKgojR07VtnZ2YqJiZHD4dDEiROVlpbGJ5YAAIC9IbNgwQJJ0k033eSzf9myZbrvvvskSXPmzFFwcLAyMzPl8XiUkZGh+fPnN/OkAAAgENkaMpZlnfeciIgI5eXlKS8vrxkmAgAAJgmIm30BAAD8QcgAAABjETIAAMBYhAwAADAWIQMAAIxFyAAAAGMRMgAAwFiEDAAAMBYhAwAAjEXIAAAAYxEyAADAWIQMAAAwFiEDAACMRcgAAABjETIAAMBYhAwAADAWIQMAAIxFyAAAAGMRMgAAwFiEDAAAMBYhAwAAjEXIAAAAYxEyAADAWIQMAAAwFiEDAACMRcgAAABjETIAAMBYhAwAADAWIQMAAIxFyAAAAGMRMgAAwFiEDAAAMBYhAwAAjEXIAAAAYxEyAADAWIQMAAAwFiEDAACMRcgAAABjETIAAMBYhAwAADAWIQMAAIxFyAAAAGMRMgAAwFi2hsz777+voUOHKiEhQUFBQVq9erXPccuy9Mwzzyg+Pl6RkZFKT0/X559/bs+wAAAg4NgaMsePH9dVV12lvLy8Oo/PmjVLL7/8shYuXKitW7fq4osvVkZGhqqqqpp5UgAAEIhC7PzDhwwZoiFDhtR5zLIszZ07V08//bSGDRsmSXr11VcVGxur1atX66677mrOUQEAQAAK2HtkDh48KKfTqfT0dO++qKgopaamqrCwsN7neTweud1unw0AALRMARsyTqdTkhQbG+uzPzY21nusLjNmzFBUVJR3S0xMbNI5AQCAfQI2ZPyVk5OjyspK71ZWVmb3SAAAoIkEbMjExcVJklwul89+l8vlPVaX8PBwORwOnw0AALRMARsyycnJiouLU0FBgXef2+3W1q1blZaWZuNkAAAgUNj6qaVjx45p//793scHDx7Uzp07FRMTo6SkJE2aNEnPP/+8Lr/8ciUnJ2vKlClKSEjQ8OHD7RsaAAAEDFtDZvv27Ro0aJD3cXZ2tiRpzJgxWr58uZ544gkdP35c48ePV0VFha6//nqtW7dOERERdo0MAAACiK0hc9NNN8myrHqPBwUFadq0aZo2bVozTgUAAEwRsPfIAAAAnA8hAwAAjEXIAAAAYxEyAADAWIQMAAAwFiEDAACMRcgAAABjETIAAMBYhAwAADAWIQMAAIxFyAAAAGMRMgAAwFiEDAAAMBYhAwAAjEXIAAAAYxEyAADAWIQMAAAwFiEDAACMRcgAAABjhdg9AADg5620tFTl5eV2j9Ek9u7da/cILR4hAwCwTWlpqVJSeujkyRN2j9KkajzVdo/QYhEyAADblJeX6+TJE0r9Ta4c8Z3tHqfRHdlVqN1vLtKpU6fsHqXFImQAALZzxHdWTFJ3u8dodO4jX9o9QovHzb4AAMBYhAwAADAWIQMAAIxFyAAAAGMRMgAAwFiEDAAAMBYhAwAAjEXIAAAAYxEyAADAWIQMAAAwFiEDAACMRcgAAABj8UsjcV579+61e4Qm0759eyUlJdk9BgDAT4QM6nWy8ltJQbr33nvtHqXJREZepH379hIzAGAoQgb1qjlxVJKlq+95Upckp9g9TqNzH/lSW5dOVXl5OSEDAIYiZHBerTskKSapu91jAABwFm72BQAAxuKKDH72WvLNzJLk8XgUHh5u9xhNhvdntpb+7x+aHiGDn62fw83MkqSgIMmy7J6i6fD+WoQaT7XdI8BQRoRMXl6e/vjHP8rpdOqqq67SvHnz1L9/f7vHguFa+s3MknRkV6F2v7moxb5H3p/5zrzHU6dO2T0KDBXwIfP3v/9d2dnZWrhwoVJTUzV37lxlZGSopKREHTp0sHs8tAAt+WZm95EvJbXc98j7M9+Z9wj4K+Bv9p09e7bGjRun+++/Xz179tTChQt10UUXaenSpXaPBgAAbBbQV2Sqq6tVXFysnJwc777g4GClp6ersLCwzud4PB55PB7v48rKSkmS2+1u1NmOHTsmSfruvyU65TnZqK8dKNxH/itJqjz0uUJDgmyepvG19Pcntfz3yPszX0t/jy3+/TlLJf3wNbGxv86eeT3rfPeIWQHs0KFDliRry5YtPvsff/xxq3///nU+Jzc315LExsbGxsbG1gK2srKyc7ZCQF+R8UdOTo6ys7O9j2tra/Xdd9+pXbt2CgpqvBp2u91KTExUWVmZHA5Ho70uzsZaNw/WuXmwzs2DdW4+TbXWlmXp6NGjSkhIOOd5AR0y7du3V6tWreRyuXz2u1wuxcXF1fmc8PDws37mQnR0dFONKIfDwb8kzYS1bh6sc/NgnZsH69x8mmKto6KizntOQN/sGxYWpj59+qigoMC7r7a2VgUFBUpLS7NxMgAAEAgC+oqMJGVnZ2vMmDHq27ev+vfvr7lz5+r48eO6//777R4NAADYLOBDZuTIkfrmm2/0zDPPyOl06uqrr9a6desUGxtr61zh4eHKzc1t0T86PFCw1s2DdW4erHPzYJ2bj91rHWRZP4OffQ0AAFqkgL5HBgAA4FwIGQAAYCxCBgAAGIuQAQAAxiJkziEvL0+dO3dWRESEUlNT9dFHH53z/JUrVyolJUURERG68sortXbt2maa1HwNWevFixfrhhtuUNu2bdW2bVulp6ef938b/KChf6fPyM/PV1BQkIYPH960A7YQDV3niooKZWVlKT4+XuHh4erWrRv//bgADV3nuXPnqnv37oqMjFRiYqImT56sqqqqZprWTO+//76GDh2qhIQEBQUFafXq1ed9zqZNm3TttdcqPDxcl112mZYvX960QzbOb0VqefLz862wsDBr6dKl1qeffmqNGzfOio6OtlwuV53nf/jhh1arVq2sWbNmWXv27LGefvppKzQ01Nq1a1czT26ehq71PffcY+Xl5Vk7duyw9u7da913331WVFSU9dVXXzXz5GZp6DqfcfDgQevSSy+1brjhBmvYsGHNM6zBGrrOHo/H6tu3r3XbbbdZH3zwgXXw4EFr06ZN1s6dO5t5crM0dJ1fe+01Kzw83HrttdesgwcPWuvXr7fi4+OtyZMnN/PkZlm7dq311FNPWW+88YYlyVq1atU5zz9w4IB10UUXWdnZ2daePXusefPmWa1atbLWrVvXZDMSMvXo37+/lZWV5X18+vRpKyEhwZoxY0ad5995553W7bff7rMvNTXVevDBB5t0zpagoWv9Y6dOnbLatGljvfLKK001YovgzzqfOnXKGjBggPXXv/7VGjNmDCFzARq6zgsWLLC6dOliVVdXN9eILUJD1zkrK8saPHiwz77s7Gxr4MCBTTpnS3IhIfPEE09YvXr18tk3cuRIKyMjo8nm4ltLdaiurlZxcbHS09O9+4KDg5Wenq7CwsI6n1NYWOhzviRlZGTUez5+4M9a/9iJEydUU1OjmJiYphrTeP6u87Rp09ShQweNHTu2OcY0nj/r/OabbyotLU1ZWVmKjY3VFVdcoenTp+v06dPNNbZx/FnnAQMGqLi42PvtpwMHDmjt2rW67bbbmmXmnws7vhYG/E/2tUN5eblOnz591k8Pjo2N1b59++p8jtPprPN8p9PZZHO2BP6s9Y89+eSTSkhIOOtfHvx//qzzBx98oCVLlmjnzp3NMGHL4M86HzhwQBs3btSoUaO0du1a7d+/X4888ohqamqUm5vbHGMbx591vueee1ReXq7rr79elmXp1KlTeuihh/T73/++OUb+2ajva6Hb7dbJkycVGRnZ6H8mV2RgtJkzZyo/P1+rVq1SRESE3eO0GEePHtXo0aO1ePFitW/f3u5xWrTa2lp16NBBixYtUp8+fTRy5Eg99dRTWrhwod2jtSibNm3S9OnTNX/+fH388cd644039Pbbb+u5556zezT8RFyRqUP79u3VqlUruVwun/0ul0txcXF1PicuLq5B5+MH/qz1GS+++KJmzpypDRs2qHfv3k05pvEaus5ffPGFvvzySw0dOtS7r7a2VpIUEhKikpISde3atWmHNpA/f5/j4+MVGhqqVq1aeff16NFDTqdT1dXVCgsLa9KZTeTPOk+ZMkWjR4/WAw88IEm68sordfz4cY0fP15PPfWUgoP5//WNob6vhQ6Ho0muxkhckalTWFiY+vTpo4KCAu++2tpaFRQUKC0trc7npKWl+ZwvSe+++2695+MH/qy1JM2aNUvPPfec1q1bp759+zbHqEZr6DqnpKRo165d2rlzp3f75S9/qUGDBmnnzp1KTExszvGN4c/f54EDB2r//v3eUJSkzz77TPHx8URMPfxZ5xMnTpwVK2fi0eJXDjYaW74WNtltxIbLz8+3wsPDreXLl1t79uyxxo8fb0VHR1tOp9OyLMsaPXq09bvf/c57/ocffmiFhIRYL774orV3714rNzeXj19foIau9cyZM62wsDDrH//4h3XkyBHvdvToUbveghEaus4/xqeWLkxD17m0tNRq06aNNWHCBKukpMR66623rA4dOljPP/+8XW/BCA1d59zcXKtNmzbW3/72N+vAgQPWv/71L6tr167WnXfeaddbMMLRo0etHTt2WDt27LAkWbNnz7Z27Nhh/fe//7Usy7J+97vfWaNHj/aef+bj148//ri1d+9eKy8vj49f22nevHlWUlKSFRYWZvXv398qKiryHrvxxhutMWPG+Jz/+uuvW926dbPCwsKsXr16WW+//XYzT2yuhqx1p06dLElnbbm5uc0/uGEa+nf6fxEyF66h67xlyxYrNTXVCg8Pt7p06WK98MIL1qlTp5p5avM0ZJ1ramqsZ5991uratasVERFhJSYmWo888oj1/fffN//gBnnvvffq/O/tmbUdM2aMdeONN571nKuvvtoKCwuzunTpYi1btqxJZwyyLK6pAQAAM3GPDAAAMBYhAwAAjEXIAAAAYxEyAADAWIQMAAAwFiEDAACMRcgAAABjETIAAMBYhAwAADAWIQMAAIxFyAAAAGMRMgAAwFj/D8BHsloHhXiuAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "y_hat_proba = clf.predict_proba(X_test)[:,1]\n",
    "import seaborn as sns\n",
    "sns.histplot(y_hat_proba)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ee2e3512-474d-4e8c-9dee-c7c8d48bdcfe",
   "metadata": {},
   "source": [
    "Le modèle est assez confiant de ses prédiction, la plupart des prédictions ont une probabilité proche de 0 ou de 1"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6e7c6094-fe95-4882-ab0d-4400b1f01f30",
   "metadata": {},
   "source": [
    "## Accuracy "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "df3cb0ee-af05-4619-9b03-03fa6b83ad6d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "accuracy 0.9649122807017544\n"
     ]
    }
   ],
   "source": [
    "# predictions \n",
    "y_pred = clf.predict(X_test)\n",
    "\n",
    "from sklearn.metrics import accuracy_score\n",
    "print(\"accuracy\",accuracy_score(y_test, y_pred))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b139036e-a3e2-4193-a7da-2010472ed8ba",
   "metadata": {},
   "source": [
    "0.965 pas mal!"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "055bea14-e308-4f54-8636-a638327acc3a",
   "metadata": {},
   "source": [
    "## Matrice de confusion"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "c497b0b3-0fc2-48fa-a3cb-3c0d2a3b9603",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[40,  3],\n",
       "       [ 1, 70]])"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "confusion_matrix(y_test, y_pred)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5dbf741b-bd62-43e8-b58d-594eacdd5431",
   "metadata": {},
   "source": [
    "## Seuil de séparation des classes\n",
    "\n",
    "Comparons la matrice de confusion et les scores obtenus pour differents seuil de classification"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "df15d843-6966-4c47-adb0-6270612b1cd7",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Probabilité de la classe 1\n",
    "y_hat_proba = clf.predict_proba(X_test)[:,1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "1ed70905-72bd-49b3-a9f3-fa6fc463583a",
   "metadata": {},
   "outputs": [],
   "source": [
    "# classes prédites pour les  seuils 0.3 et 0.7\n",
    "\n",
    "y_pred_03 = [ 0 if value < 0.3 else 1 for value in y_hat_proba ]\n",
    "y_pred_07 = [ 0 if value < 0.7 else 1 for value in y_hat_proba ]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "e5e0d23f-7e5a-4559-899c-ca26832d097c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[38,  5],\n",
       "       [ 1, 70]])"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Matrice de confusion pour le seuil 0.3\n",
    "\n",
    "confusion_matrix(y_test, y_pred_03)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "f348bd58-2892-4e58-ab22-7685d10edcdc",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[41,  2],\n",
       "       [ 2, 69]])"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Matrice de confusion pour le seuil 0.7\n",
    "\n",
    "confusion_matrix(y_test, y_pred_07)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3c0537b1-d8b5-4a16-882c-1289be946a3a",
   "metadata": {},
   "source": [
    "## Autres metriques de classification\n",
    "Precision, Recall et ROC_AUC\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "e8dece65-956c-4d5a-bcb9-6d05bab28bfa",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Precision: 0.958904109589041\n",
      "Recall: 0.9859154929577465\n",
      "ROC-AUC 0.9967245332459875\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import precision_score, recall_score, roc_auc_score\n",
    "print(\"Precision:\",precision_score(y_test, y_pred))\n",
    "print(\"Recall:\",recall_score(y_test, y_pred))\n",
    "print(\"ROC-AUC\", roc_auc_score(y_test, clf.predict_proba(X_test)[:, 1]))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ebbad5fd-191b-44be-80d4-17f77e5d14ef",
   "metadata": {},
   "source": [
    "### Courbe ROC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "2642771f-acd4-414e-8187-f39c29bcf8ff",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'ROC curve')"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import roc_curve\n",
    "fpr, tpr, thresholds = roc_curve(y_test, clf.predict_proba(X_test)[:,1])\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "plt.plot(fpr, tpr)\n",
    "plt.grid()\n",
    "plt.title(\"ROC curve\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2ca498fd-82f0-4d5a-899d-90778653b9c9",
   "metadata": {},
   "source": [
    "## Classification sur le dataset IRIS\n",
    "\n",
    "Le dataset [IRIS](https://scikit-learn.org/stable/auto_examples/datasets/plot_iris_dataset.html) est disponible dans scikit-learn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "3b937b6f-2be3-4c1b-8038-11f9cfe82504",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.datasets import load_iris\n",
    "X, y = load_iris(return_X_y=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "id": "0da78399-de53-41d1-9aa1-b3d3c56bbc0c",
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.20, random_state=8)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "id": "0e74ddc6-0eb1-4e82-a6fd-64b044cac788",
   "metadata": {},
   "outputs": [],
   "source": [
    "# entrainons le modele\n",
    "\n",
    "clf = LogisticRegression(random_state=8).fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "id": "e9ec2c68-3b37-4098-b7f4-c404b1688702",
   "metadata": {},
   "outputs": [],
   "source": [
    "# prediction sur le set de test\n",
    "\n",
    "y_pred = clf.predict(X_test)\n",
    "y_pred_proba = clf.predict_proba(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "id": "a96a06f7-6785-4e6e-9cab-026a471d28d2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[10,  0,  0],\n",
       "       [ 0,  8,  1],\n",
       "       [ 0,  2,  9]])"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Matrice de confusion\n",
    "confusion_matrix(y_test, y_pred)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ed5893b3-e802-44fe-b322-90763235c6bb",
   "metadata": {},
   "source": [
    "Comme il y a  3 classes à prédire, la matrice de confusion est une matric 3x3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "id": "6263f550-6b77-40f8-9cbb-c3ee7f46cceb",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ROC-AUC 0.9832071761896323\n"
     ]
    }
   ],
   "source": [
    "\n",
    "print(\"ROC-AUC\", roc_auc_score(y_test, clf.predict_proba(X_test),multi_class='ovr'))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "26d3f2e4-1701-461a-a3ab-d01fae16e132",
   "metadata": {},
   "source": [
    "##  classification report\n",
    "Le classification report donne differentes metriques pour chacune des classes."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "id": "95393f70-3421-4d79-a2e6-fbbffcb9b8e2",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      1.00      1.00        10\n",
      "           1       0.80      0.89      0.84         9\n",
      "           2       0.90      0.82      0.86        11\n",
      "\n",
      "    accuracy                           0.90        30\n",
      "   macro avg       0.90      0.90      0.90        30\n",
      "weighted avg       0.90      0.90      0.90        30\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import classification_report\n",
    "\n",
    "print(classification_report(y_test, y_pred))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "bab75cee-1046-472f-9a13-d77e68039007",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.11.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}