Kaplan-Meier.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "from lifelines import KaplanMeierFitter\n",
    "import pandas as pd\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# サンプルデータの生成\n",
    "# ここでは、`duration`は顧客がサービスを利用していた期間（例えば、月数）、\n",
    "# `event_observed`は顧客が解約したかどうか（1は解約、0はまだ継続中または右側打ち切りデータ）を示します。\n",
    "data = {\n",
    "    'duration': [5, 6, 6, 2.5, 4, 3],  # サービス利用期間\n",
    "    'event_observed': [1, 0, 1, 1, 1, 0]  # 解約したかどうか（1: はい、0: いいえ）\n",
    "}\n",
    "df = pd.DataFrame(data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "          KM_estimate\n",
      "timeline             \n",
      "0.0          1.000000\n",
      "2.5          0.833333\n",
      "3.0          0.833333\n",
      "4.0          0.625000\n",
      "5.0          0.416667\n",
      "6.0          0.208333\n",
      "5.0\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Kaplan-Meierフィッターのインスタンス化\n",
    "kmf = KaplanMeierFitter()\n",
    "\n",
    "# データフィット\n",
    "kmf.fit(durations=df['duration'], event_observed=df['event_observed'])\n",
    "\n",
    "# 生存曲線のプロット\n",
    "kmf.plot_survival_function()\n",
    "\n",
    "# 結果の表示\n",
    "print(kmf.survival_function_)\n",
    "print(kmf.median_survival_time_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}