{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "73f04560",
   "metadata": {},
   "source": [
    "# 第 08 章　K-平均分群（K-Means Clustering）\n",
    "\n",
    "本 notebook 搭配網站「醫學生的機器學習入門」第 08 章。建議在 Google Colab 開啟：上方選單「執行階段 → 全部執行」即可從頭跑完（約 1 分鐘，不需要 GPU）。\n",
    "\n",
    "這一章是本課程第一個**非監督學習（unsupervised learning）**：資料沒有答案欄位，我們要讓演算法自己找出「哪些病人彼此比較像」。\n",
    "\n",
    "流程：\n",
    "1. 用 2D 玩具資料，徒手寫出 K-Means 的四個步驟\n",
    "2. 用 scikit-learn 的 `KMeans` 重做一次\n",
    "3. 用 Heart Failure Clinical Records 資料，把心衰竭病人分群，再**事後**比較各群死亡比例\n",
    "4. 看 K-Means 會在哪裡失敗\n",
    "\n",
    "> 圖上的標籤用英文，是因為 Colab 預設沒有中文字型（中文會變成方塊）。"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "51a174cb",
   "metadata": {},
   "source": [
    "## 0. 環境與版本\n",
    "\n",
    "本章只用 Colab 預裝的套件，不需要另外安裝。先印出版本，之後若遇到錯誤比較好對照。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "fde34d48",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-09-29T20:02:27.550407Z",
     "iopub.status.busy": "2026-09-29T20:02:27.550293Z",
     "iopub.status.idle": "2026-09-29T20:02:28.459438Z",
     "shell.execute_reply": "2026-09-29T20:02:28.459089Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Python       3.12.2\n",
      "numpy        2.1.3\n",
      "pandas       2.2.3\n",
      "scikit-learn 1.6.1\n"
     ]
    }
   ],
   "source": [
    "import sys\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import sklearn\n",
    "from sklearn.cluster import KMeans\n",
    "from sklearn.datasets import make_blobs, make_moons\n",
    "from sklearn.metrics import silhouette_score, adjusted_rand_score\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "\n",
    "# Some numpy 2.x builds on macOS (Apple Accelerate) print spurious\n",
    "# \"encountered in matmul\" RuntimeWarnings inside KMeans; results are unaffected.\n",
    "# Colab (Linux) does not show them. Silence only that message, only on macOS.\n",
    "if sys.platform == \"darwin\":\n",
    "    import warnings\n",
    "    warnings.filterwarnings(\"ignore\", message=\".*encountered in matmul\", category=RuntimeWarning)\n",
    "\n",
    "print(\"Python      \", sys.version.split()[0])\n",
    "print(\"numpy       \", np.__version__)\n",
    "print(\"pandas      \", pd.__version__)\n",
    "print(\"scikit-learn\", sklearn.__version__)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b199a42c",
   "metadata": {},
   "source": [
    "## 1. 徒手寫 K-Means：四個步驟\n",
    "\n",
    "先做一份「答案很明顯」的玩具資料：300 個點，其實來自 3 團。我們假裝不知道它們屬於哪一團。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "99ae4b4e",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-09-29T20:02:28.460673Z",
     "iopub.status.busy": "2026-09-29T20:02:28.460574Z",
     "iopub.status.idle": "2026-09-29T20:02:28.504142Z",
     "shell.execute_reply": "2026-09-29T20:02:28.503714Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(300, 2)\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 450x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "X_toy, _ = make_blobs(n_samples=300, centers=3, cluster_std=0.9, random_state=42)\n",
    "print(X_toy.shape)\n",
    "plt.figure(figsize=(4.5, 4))\n",
    "plt.scatter(X_toy[:, 0], X_toy[:, 1], s=12, color=\"gray\")\n",
    "plt.title(\"Toy data (no labels)\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c766714d",
   "metadata": {},
   "source": [
    "下面兩個函式就是 K-Means 的全部核心：\n",
    "\n",
    "- `assign`（分配）：每個點找距離最近的群中心（centroid）\n",
    "- `update`（更新）：每個群中心搬到自己成員的平均位置\n",
    "\n",
    "只要反覆做這兩步，直到群中心不再移動。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "2ae2aedf",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-09-29T20:02:28.505443Z",
     "iopub.status.busy": "2026-09-29T20:02:28.505368Z",
     "iopub.status.idle": "2026-09-29T20:02:28.609870Z",
     "shell.execute_reply": "2026-09-29T20:02:28.609489Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "iteration 1: centers moved 6.620\n",
      "iteration 2: centers moved 0.376\n",
      "iteration 3: centers moved 0.497\n",
      "iteration 4: centers moved 1.106\n",
      "iteration 5: centers moved 5.433\n",
      "iteration 6: centers moved 5.937\n",
      "iteration 7: centers moved 0.000\n",
      "converged, cluster sizes: [100 100 100]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1400x340 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "def assign(X, centers):\n",
    "    # distance from every point to every center: shape (n_points, k)\n",
    "    d = np.linalg.norm(X[:, None, :] - centers[None, :, :], axis=2)\n",
    "    return d.argmin(axis=1)\n",
    "\n",
    "def update(X, labels, k):\n",
    "    return np.array([X[labels == j].mean(axis=0) for j in range(k)])\n",
    "\n",
    "rng = np.random.default_rng(6)\n",
    "k = 3\n",
    "centers = X_toy[rng.choice(len(X_toy), size=k, replace=False)]   # step 1: pick k random points as initial centers\n",
    "\n",
    "snapshots = []\n",
    "for it in range(1, 21):\n",
    "    labels = assign(X_toy, centers)                                 # step 2: assign\n",
    "    snapshots.append((it, labels, centers))\n",
    "    new_centers = update(X_toy, labels, k)                          # step 3: update\n",
    "    shift = np.linalg.norm(new_centers - centers)\n",
    "    print(f\"iteration {it}: centers moved {shift:.3f}\")\n",
    "    centers = new_centers\n",
    "    if shift < 1e-6:                                                # step 4: repeat until nothing moves\n",
    "        print(\"converged, cluster sizes:\", np.bincount(labels))\n",
    "        break\n",
    "\n",
    "show = snapshots[:3] + [snapshots[-1]]\n",
    "fig, axes = plt.subplots(1, 4, figsize=(14, 3.4), sharex=True, sharey=True)\n",
    "for ax, (it, labels, c) in zip(axes, show):\n",
    "    ax.scatter(X_toy[:, 0], X_toy[:, 1], c=labels, cmap=\"tab10\", vmin=0, vmax=9, s=10)\n",
    "    ax.scatter(c[:, 0], c[:, 1], c=\"black\", marker=\"X\", s=150)\n",
    "    ax.set_title(f\"Iteration {it}\")\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "24605499",
   "metadata": {},
   "source": [
    "前幾輪群中心會大幅移動，有時某個中心還會「跳」到另一團去；當移動距離變成 0，代表分配不再改變，也就是收斂了。最後每群 100 點，正好是資料產生時的三團。"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bab47e58",
   "metadata": {},
   "source": [
    "## 2. 用 scikit-learn 做同一件事\n",
    "\n",
    "實務上不用自己寫。`KMeans` 預設用 k-means++ 挑選比較分散的初始中心；`n_init=10` 代表用 10 組不同初始值各跑一次，留下最好的那次；`random_state` 固定讓結果可重現。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "c79b1de4",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-09-29T20:02:28.610935Z",
     "iopub.status.busy": "2026-09-29T20:02:28.610868Z",
     "iopub.status.idle": "2026-09-29T20:02:28.641664Z",
     "shell.execute_reply": "2026-09-29T20:02:28.641070Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "cluster sizes: [100 100 100]\n",
      "inertia (within-cluster sum of squares): 459.2\n",
      "silhouette: 0.863\n"
     ]
    }
   ],
   "source": [
    "km_toy = KMeans(n_clusters=3, n_init=10, random_state=42).fit(X_toy)\n",
    "print(\"cluster sizes:\", np.bincount(km_toy.labels_))\n",
    "print(\"inertia (within-cluster sum of squares):\", round(km_toy.inertia_, 1))\n",
    "print(\"silhouette:\", round(silhouette_score(X_toy, km_toy.labels_), 3))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "53484b36",
   "metadata": {},
   "source": [
    "`inertia_` 是「每個點到自己群中心距離平方的總和」，越小代表群越緊。輪廓係數（silhouette）介於 -1 到 1，越接近 1 代表群內緊、群間分得開；玩具資料的值會很高。記住這個數字，等一下跟真實病人資料比較。"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "32435381",
   "metadata": {},
   "source": [
    "## 3. 真實資料：Heart Failure Clinical Records\n",
    "\n",
    "299 位心衰竭住院病人（巴基斯坦 Faisalabad 兩家醫院，2015 年），來源 UCI Machine Learning Repository（ID 519，CC BY 4.0）。原始研究：Ahmad T, et al. *PLoS One* 2017;12:e0181001；Chicco D, Jurman G. *BMC Med Inform Decis Mak* 2020;20:16。\n",
    "\n",
    "先試 UCI 的 CSV 直連；失敗再改用 `ucimlrepo`（Colab 需先 `%pip install ucimlrepo`）。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "99a4f78e",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-09-29T20:02:28.643071Z",
     "iopub.status.busy": "2026-09-29T20:02:28.643002Z",
     "iopub.status.idle": "2026-09-29T20:02:29.137655Z",
     "shell.execute_reply": "2026-09-29T20:02:29.136820Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(299, 13)\n"
     ]
    },
    {
     "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>age</th>\n",
       "      <th>anaemia</th>\n",
       "      <th>creatinine_phosphokinase</th>\n",
       "      <th>diabetes</th>\n",
       "      <th>ejection_fraction</th>\n",
       "      <th>high_blood_pressure</th>\n",
       "      <th>platelets</th>\n",
       "      <th>serum_creatinine</th>\n",
       "      <th>serum_sodium</th>\n",
       "      <th>sex</th>\n",
       "      <th>smoking</th>\n",
       "      <th>time</th>\n",
       "      <th>death_event</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>75.0</td>\n",
       "      <td>0</td>\n",
       "      <td>582</td>\n",
       "      <td>0</td>\n",
       "      <td>20</td>\n",
       "      <td>1</td>\n",
       "      <td>265000.00</td>\n",
       "      <td>1.9</td>\n",
       "      <td>130</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>55.0</td>\n",
       "      <td>0</td>\n",
       "      <td>7861</td>\n",
       "      <td>0</td>\n",
       "      <td>38</td>\n",
       "      <td>0</td>\n",
       "      <td>263358.03</td>\n",
       "      <td>1.1</td>\n",
       "      <td>136</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>6</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>65.0</td>\n",
       "      <td>0</td>\n",
       "      <td>146</td>\n",
       "      <td>0</td>\n",
       "      <td>20</td>\n",
       "      <td>0</td>\n",
       "      <td>162000.00</td>\n",
       "      <td>1.3</td>\n",
       "      <td>129</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>7</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>50.0</td>\n",
       "      <td>1</td>\n",
       "      <td>111</td>\n",
       "      <td>0</td>\n",
       "      <td>20</td>\n",
       "      <td>0</td>\n",
       "      <td>210000.00</td>\n",
       "      <td>1.9</td>\n",
       "      <td>137</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>7</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>65.0</td>\n",
       "      <td>1</td>\n",
       "      <td>160</td>\n",
       "      <td>1</td>\n",
       "      <td>20</td>\n",
       "      <td>0</td>\n",
       "      <td>327000.00</td>\n",
       "      <td>2.7</td>\n",
       "      <td>116</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>8</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    age  anaemia  creatinine_phosphokinase  diabetes  ejection_fraction  \\\n",
       "0  75.0        0                       582         0                 20   \n",
       "1  55.0        0                      7861         0                 38   \n",
       "2  65.0        0                       146         0                 20   \n",
       "3  50.0        1                       111         0                 20   \n",
       "4  65.0        1                       160         1                 20   \n",
       "\n",
       "   high_blood_pressure  platelets  serum_creatinine  serum_sodium  sex  \\\n",
       "0                    1  265000.00               1.9           130    1   \n",
       "1                    0  263358.03               1.1           136    1   \n",
       "2                    0  162000.00               1.3           129    1   \n",
       "3                    0  210000.00               1.9           137    1   \n",
       "4                    0  327000.00               2.7           116    0   \n",
       "\n",
       "   smoking  time  death_event  \n",
       "0        0     4            1  \n",
       "1        0     6            1  \n",
       "2        1     7            1  \n",
       "3        0     7            1  \n",
       "4        0     8            1  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "URL = (\"https://archive.ics.uci.edu/ml/machine-learning-databases/00519/\"\n",
    "       \"heart_failure_clinical_records_dataset.csv\")\n",
    "try:\n",
    "    df = pd.read_csv(URL)\n",
    "except Exception as e:\n",
    "    print(\"Direct CSV download failed:\", e)\n",
    "    try:\n",
    "        from ucimlrepo import fetch_ucirepo\n",
    "        r = fetch_ucirepo(id=519)\n",
    "        df = r.data.features.join(r.data.targets)\n",
    "    except Exception as e2:\n",
    "        raise RuntimeError(\n",
    "            \"Cannot download the Heart Failure dataset. Check the internet connection, \"\n",
    "            \"or run `%pip install ucimlrepo` and retry.\"\n",
    "        ) from e2\n",
    "df.columns = df.columns.str.lower()\n",
    "print(df.shape)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6ca356f2",
   "metadata": {},
   "source": [
    "我們只拿 6 個**連續變數**來分群，`death_event`（追蹤期間是否死亡）先藏起來，最後才拿出來對照。二元變數（貧血、糖尿病、性別⋯⋯）只有 0/1 兩個值，用「距離」衡量不太合理，本章先不放進去。`time`（追蹤天數）跟結果綁在一起，也不放。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "d17c598d",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-09-29T20:02:29.139571Z",
     "iopub.status.busy": "2026-09-29T20:02:29.139433Z",
     "iopub.status.idle": "2026-09-29T20:02:29.150275Z",
     "shell.execute_reply": "2026-09-29T20:02:29.149873Z"
    }
   },
   "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>count</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>min</th>\n",
       "      <th>25%</th>\n",
       "      <th>50%</th>\n",
       "      <th>75%</th>\n",
       "      <th>max</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>age</th>\n",
       "      <td>299.0</td>\n",
       "      <td>60.83</td>\n",
       "      <td>11.89</td>\n",
       "      <td>40.0</td>\n",
       "      <td>51.0</td>\n",
       "      <td>60.0</td>\n",
       "      <td>70.0</td>\n",
       "      <td>95.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>creatinine_phosphokinase</th>\n",
       "      <td>299.0</td>\n",
       "      <td>581.84</td>\n",
       "      <td>970.29</td>\n",
       "      <td>23.0</td>\n",
       "      <td>116.5</td>\n",
       "      <td>250.0</td>\n",
       "      <td>582.0</td>\n",
       "      <td>7861.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ejection_fraction</th>\n",
       "      <td>299.0</td>\n",
       "      <td>38.08</td>\n",
       "      <td>11.83</td>\n",
       "      <td>14.0</td>\n",
       "      <td>30.0</td>\n",
       "      <td>38.0</td>\n",
       "      <td>45.0</td>\n",
       "      <td>80.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>platelets</th>\n",
       "      <td>299.0</td>\n",
       "      <td>263358.03</td>\n",
       "      <td>97804.24</td>\n",
       "      <td>25100.0</td>\n",
       "      <td>212500.0</td>\n",
       "      <td>262000.0</td>\n",
       "      <td>303500.0</td>\n",
       "      <td>850000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>serum_creatinine</th>\n",
       "      <td>299.0</td>\n",
       "      <td>1.39</td>\n",
       "      <td>1.03</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0.9</td>\n",
       "      <td>1.1</td>\n",
       "      <td>1.4</td>\n",
       "      <td>9.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>serum_sodium</th>\n",
       "      <td>299.0</td>\n",
       "      <td>136.63</td>\n",
       "      <td>4.41</td>\n",
       "      <td>113.0</td>\n",
       "      <td>134.0</td>\n",
       "      <td>137.0</td>\n",
       "      <td>140.0</td>\n",
       "      <td>148.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                          count       mean       std      min       25%  \\\n",
       "age                       299.0      60.83     11.89     40.0      51.0   \n",
       "creatinine_phosphokinase  299.0     581.84    970.29     23.0     116.5   \n",
       "ejection_fraction         299.0      38.08     11.83     14.0      30.0   \n",
       "platelets                 299.0  263358.03  97804.24  25100.0  212500.0   \n",
       "serum_creatinine          299.0       1.39      1.03      0.5       0.9   \n",
       "serum_sodium              299.0     136.63      4.41    113.0     134.0   \n",
       "\n",
       "                               50%       75%       max  \n",
       "age                           60.0      70.0      95.0  \n",
       "creatinine_phosphokinase     250.0     582.0    7861.0  \n",
       "ejection_fraction             38.0      45.0      80.0  \n",
       "platelets                 262000.0  303500.0  850000.0  \n",
       "serum_creatinine               1.1       1.4       9.4  \n",
       "serum_sodium                 137.0     140.0     148.0  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cols = [\"age\", \"creatinine_phosphokinase\", \"ejection_fraction\",\n",
    "        \"platelets\", \"serum_creatinine\", \"serum_sodium\"]\n",
    "df[cols].describe().round(2).T"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2c7a64c1",
   "metadata": {},
   "source": [
    "注意兩件事：\n",
    "1. 各變數的**單位與數量級差很多**：platelets 是幾十萬，serum_creatinine 只有 1 左右。\n",
    "2. creatinine_phosphokinase（CPK）與 serum_creatinine **右偏很嚴重**（最大值是中位數的好幾倍）。\n",
    "\n",
    "### 3.1 故意做錯：不標準化直接分群\n",
    "\n",
    "K-Means 用的是歐氏距離，數字大的變數會主導距離。先看看直接丟原始數值會怎樣。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "44a48b5e",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-09-29T20:02:29.151698Z",
     "iopub.status.busy": "2026-09-29T20:02:29.151616Z",
     "iopub.status.idle": "2026-09-29T20:02:29.162941Z",
     "shell.execute_reply": "2026-09-29T20:02:29.162267Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "         size       min       max\n",
      "cluster                          \n",
      "0         184  210000.0  338000.0\n",
      "1          45  348000.0  850000.0\n",
      "2          70   25100.0  208000.0\n"
     ]
    }
   ],
   "source": [
    "km_raw = KMeans(n_clusters=3, n_init=10, random_state=42).fit(df[cols])\n",
    "print(df.assign(cluster=km_raw.labels_).groupby(\"cluster\")[\"platelets\"]\n",
    "        .agg([\"size\", \"min\", \"max\"]))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "df80c6fe",
   "metadata": {},
   "source": [
    "三群的血小板範圍幾乎**完全不重疊**：演算法其實只是按血小板數切成低、中、高三段，其他五個變數幾乎沒發言權。這不是醫學發現，而是單位造成的假象。\n",
    "\n",
    "### 3.2 正確做法：偏態變數取對數，再標準化\n",
    "\n",
    "CPK 與肌酸酐先取 log 壓縮長尾（減少極端值拉走群中心），再用 `StandardScaler` 把每個變數轉成平均 0、標準差 1。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "b1ef4b08",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-09-29T20:02:29.164592Z",
     "iopub.status.busy": "2026-09-29T20:02:29.164491Z",
     "iopub.status.idle": "2026-09-29T20:02:29.168573Z",
     "shell.execute_reply": "2026-09-29T20:02:29.167891Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0. 0. 0. 0. 0. 0.]\n",
      "[1. 1. 1. 1. 1. 1.]\n"
     ]
    }
   ],
   "source": [
    "X = df[cols].copy()\n",
    "X[\"creatinine_phosphokinase\"] = np.log(X[\"creatinine_phosphokinase\"])\n",
    "X[\"serum_creatinine\"] = np.log(X[\"serum_creatinine\"])\n",
    "X_scaled = StandardScaler().fit_transform(X)\n",
    "print(X_scaled.mean(axis=0).round(2) + 0.0)   # + 0.0 turns -0.0 into 0.0\n",
    "print(X_scaled.std(axis=0).round(2))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5418fc72",
   "metadata": {},
   "source": [
    "每一欄平均都變成 0、標準差都變成 1，六個變數現在站在同一個起跑點。\n",
    "\n",
    "### 3.3 要分幾群？手肘法與輪廓係數\n",
    "\n",
    "K 要事先指定。把 K 從 2 試到 8，記錄 inertia（手肘法，elbow method）與平均輪廓係數。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "c83d35f1",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-09-29T20:02:29.169816Z",
     "iopub.status.busy": "2026-09-29T20:02:29.169756Z",
     "iopub.status.idle": "2026-09-29T20:02:29.286303Z",
     "shell.execute_reply": "2026-09-29T20:02:29.285751Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " k  inertia  silhouette\n",
      " 2   1514.4       0.181\n",
      " 3   1336.3       0.152\n",
      " 4   1208.8       0.137\n",
      " 5   1104.7       0.147\n",
      " 6   1021.2       0.145\n",
      " 7    947.0       0.160\n",
      " 8    899.2       0.156\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x360 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "ks = range(2, 9)\n",
    "inertias, sils = [], []\n",
    "for k in ks:\n",
    "    km = KMeans(n_clusters=k, n_init=10, random_state=42).fit(X_scaled)\n",
    "    inertias.append(km.inertia_)\n",
    "    sils.append(silhouette_score(X_scaled, km.labels_))\n",
    "\n",
    "print(pd.DataFrame({\"k\": list(ks),\n",
    "                    \"inertia\": [round(v, 1) for v in inertias],\n",
    "                    \"silhouette\": [round(v, 3) for v in sils]}).to_string(index=False))\n",
    "\n",
    "fig, axes = plt.subplots(1, 2, figsize=(10, 3.6))\n",
    "axes[0].plot(list(ks), inertias, \"o-\")\n",
    "axes[0].set_xlabel(\"k\"); axes[0].set_ylabel(\"inertia\"); axes[0].set_title(\"Elbow method\")\n",
    "axes[1].plot(list(ks), sils, \"o-\", color=\"tab:orange\")\n",
    "axes[1].set_xlabel(\"k\"); axes[1].set_ylabel(\"mean silhouette\"); axes[1].set_title(\"Silhouette\")\n",
    "axes[1].set_ylim(0, 1)\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b182cc2d",
   "metadata": {},
   "source": [
    "跟玩具資料比，這裡的手肘**不明顯**，輪廓係數在每個 K 都低於 0.2——代表這群病人並沒有「涇渭分明的天然群」，比較像一片連續分布。這在真實臨床資料很常見。\n",
    "\n",
    "K=2 的輪廓係數最高，但差距很小；為了示範多一點亞群的樣貌，下面用 K=3（K=2 留給你當練習）。這是一個**需要說清楚理由的主觀選擇**，不是演算法算出的標準答案。"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "00a00f8b",
   "metadata": {},
   "source": [
    "### 3.4 K=3 分群與各群輪廓"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "8e0b536d",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-09-29T20:02:29.287826Z",
     "iopub.status.busy": "2026-09-29T20:02:29.287734Z",
     "iopub.status.idle": "2026-09-29T20:02:29.299780Z",
     "shell.execute_reply": "2026-09-29T20:02:29.299301Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "cluster sizes: {0: 69, 1: 151, 2: 79}\n"
     ]
    },
    {
     "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>age</th>\n",
       "      <th>creatinine_phosphokinase</th>\n",
       "      <th>ejection_fraction</th>\n",
       "      <th>platelets</th>\n",
       "      <th>serum_creatinine</th>\n",
       "      <th>serum_sodium</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>cluster</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>65.0</td>\n",
       "      <td>246.0</td>\n",
       "      <td>30.0</td>\n",
       "      <td>243000.00</td>\n",
       "      <td>1.83</td>\n",
       "      <td>134.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>53.0</td>\n",
       "      <td>478.0</td>\n",
       "      <td>35.0</td>\n",
       "      <td>263358.03</td>\n",
       "      <td>1.00</td>\n",
       "      <td>138.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>65.0</td>\n",
       "      <td>129.0</td>\n",
       "      <td>50.0</td>\n",
       "      <td>255000.00</td>\n",
       "      <td>1.10</td>\n",
       "      <td>138.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          age  creatinine_phosphokinase  ejection_fraction  platelets  \\\n",
       "cluster                                                                 \n",
       "0        65.0                     246.0               30.0  243000.00   \n",
       "1        53.0                     478.0               35.0  263358.03   \n",
       "2        65.0                     129.0               50.0  255000.00   \n",
       "\n",
       "         serum_creatinine  serum_sodium  \n",
       "cluster                                  \n",
       "0                    1.83         134.0  \n",
       "1                    1.00         138.0  \n",
       "2                    1.10         138.0  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "km = KMeans(n_clusters=3, n_init=10, random_state=42).fit(X_scaled)\n",
    "df[\"cluster\"] = km.labels_\n",
    "print(\"cluster sizes:\", df[\"cluster\"].value_counts().sort_index().to_dict())\n",
    "profile = df.groupby(\"cluster\")[cols].median().round(2)\n",
    "profile"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5002b6e6",
   "metadata": {},
   "source": [
    "用**原始單位的中位數**描述各群，比看標準化後的數字容易解讀。找找看：哪一群年紀較大、肌酸酐較高、血鈉較低、射出分率較低？\n",
    "\n",
    "### 3.5 事後對照：各群死亡比例\n",
    "\n",
    "分群時完全沒有用到 `death_event`。現在才把它拿出來，看各群在追蹤期間的死亡比例。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "a9f0a23a",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-09-29T20:02:29.301046Z",
     "iopub.status.busy": "2026-09-29T20:02:29.300990Z",
     "iopub.status.idle": "2026-09-29T20:02:29.304334Z",
     "shell.execute_reply": "2026-09-29T20:02:29.303814Z"
    }
   },
   "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>n</th>\n",
       "      <th>deaths</th>\n",
       "      <th>death_rate</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>cluster</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>69</td>\n",
       "      <td>44</td>\n",
       "      <td>0.638</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>151</td>\n",
       "      <td>35</td>\n",
       "      <td>0.232</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>79</td>\n",
       "      <td>17</td>\n",
       "      <td>0.215</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           n  deaths  death_rate\n",
       "cluster                         \n",
       "0         69      44       0.638\n",
       "1        151      35       0.232\n",
       "2         79      17       0.215"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "outcome = df.groupby(\"cluster\")[\"death_event\"].agg(n=\"size\", deaths=\"sum\", death_rate=\"mean\")\n",
    "outcome[\"death_rate\"] = outcome[\"death_rate\"].round(3)\n",
    "outcome"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "94883d8b",
   "metadata": {},
   "source": [
    "這是**描述性**的比較：\n",
    "- 各群本來就是依肌酸酐、射出分率等變數劃分的，而這些變數本身就和預後有關，看到死亡比例不同並不意外。\n",
    "- 這是巴基斯坦 Faisalabad 兩家醫院、299 人的觀察性資料，**不能**解讀為「屬於某群導致死亡」，也不代表這三群是真實存在的疾病亞型。\n",
    "- 要主張某種分群有臨床價值，需要在其他醫院的資料重現（外部驗證），並確認它比既有的評估方式提供更多資訊。\n",
    "\n",
    "本例僅供學習，不構成臨床建議。\n",
    "\n",
    "### 3.6 分群穩不穩？換個亂數種子試試\n",
    "\n",
    "用 `n_init=1`（只跑一組初始值）與 `n_init=10`，各換 10 個 `random_state`，用調整蘭德指數（adjusted Rand index, ARI；1 = 分群完全相同）比較結果和第一次有多像。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "e3d3750c",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-09-29T20:02:29.305658Z",
     "iopub.status.busy": "2026-09-29T20:02:29.305585Z",
     "iopub.status.idle": "2026-09-29T20:02:29.386480Z",
     "shell.execute_reply": "2026-09-29T20:02:29.385943Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "n_init= 1: ARI vs first run  min=0.22  median=0.47; inertia range 1336.0-1418.9\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "n_init=10: ARI vs first run  min=0.46  median=0.94; inertia range 1335.7-1337.8\n"
     ]
    }
   ],
   "source": [
    "for n_init in [1, 10]:\n",
    "    runs = [KMeans(n_clusters=3, n_init=n_init, random_state=s).fit(X_scaled) for s in range(10)]\n",
    "    ari = [adjusted_rand_score(runs[0].labels_, r.labels_) for r in runs[1:]]\n",
    "    inert = [r.inertia_ for r in runs]\n",
    "    print(f\"n_init={n_init:2d}: ARI vs first run  min={min(ari):.2f}  median={np.median(ari):.2f}; \"\n",
    "          f\"inertia range {min(inert):.1f}-{max(inert):.1f}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0749bccf",
   "metadata": {},
   "source": [
    "只跑一組初始值時，換個種子結果就可能差很多；`n_init=10` 會穩定許多，但在結構不明顯的資料上仍不一定完全相同。報告分群結果時，應該附上穩定性檢查。\n",
    "\n",
    "## 4. K-Means 的限制：非球形的群\n",
    "\n",
    "K-Means 假設每一群大致是「圓的一團」。兩個互相咬合的半月形就會分錯。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "155fac64",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-09-29T20:02:29.387807Z",
     "iopub.status.busy": "2026-09-29T20:02:29.387738Z",
     "iopub.status.idle": "2026-09-29T20:02:29.456950Z",
     "shell.execute_reply": "2026-09-29T20:02:29.456460Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "agreement with true shapes (ARI): 0.25\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 900x360 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "X_moon, y_moon = make_moons(n_samples=300, noise=0.06, random_state=42)\n",
    "km_moon = KMeans(n_clusters=2, n_init=10, random_state=42).fit(X_moon)\n",
    "print(\"agreement with true shapes (ARI):\", round(adjusted_rand_score(y_moon, km_moon.labels_), 2))\n",
    "\n",
    "fig, axes = plt.subplots(1, 2, figsize=(9, 3.6), sharey=True)\n",
    "axes[0].scatter(X_moon[:, 0], X_moon[:, 1], c=y_moon, cmap=\"tab10\", vmin=0, vmax=9, s=10)\n",
    "axes[0].set_title(\"True groups\")\n",
    "axes[1].scatter(X_moon[:, 0], X_moon[:, 1], c=km_moon.labels_, cmap=\"tab10\", vmin=0, vmax=9, s=10)\n",
    "axes[1].scatter(*km_moon.cluster_centers_.T, c=\"black\", marker=\"X\", s=150)\n",
    "axes[1].set_title(\"K-Means (k=2)\")\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6ae2c23a",
   "metadata": {},
   "source": [
    "K-Means 把平面從中間切一刀，而不是沿著兩個月牙的形狀分。這類資料要改用 DBSCAN 或階層式分群（hierarchical clustering）等方法。\n",
    "\n",
    "## 5. 動手試試\n",
    "\n",
    "1. 把 3.4 的 `n_clusters=3` 改成 `2`，重跑 3.4 與 3.5：兩群的樣貌與死亡比例怎麼變？\n",
    "2. 在 3.2 拿掉取對數那兩行（只標準化），重跑 3.3 到 3.5：輪廓係數與各群輪廓有什麼改變？\n",
    "3. 在第 1 節把 `np.random.default_rng(6)` 的 6 改成 0，重跑：最後各群人數是不是還是 100／100／100？這就是「初始值不好、卡在局部最佳解」。再和第 2 節 `KMeans(n_init=10)` 的結果比較。"
   ]
  }
 ],
 "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": "3.12.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}
