Snippet to calculate k-means

k-means.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.cluster import KMeans\n",
    "from sklearn import metrics\n",
    "from sklearn.metrics import pairwise_distances\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "np.set_printoptions(threshold=np.nan)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv(\"all_ligand_descriptors.csv\")\n",
    "features = df.iloc[:, 2:]\n",
    "\n",
    "scaler = StandardScaler()\n",
    "scaler.fit(features)\n",
    "features_scaled = scaler.transform(features)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "def calc_kmeans(n_clusters, features):\n",
    "    kmeans = KMeans(n_clusters, random_state=13, n_jobs=-1)\n",
    "    kmeans.fit(features)\n",
    "    cluster_assignments = kmeans.labels_\n",
    "    cluster_centers = kmeans.cluster_centers_\n",
    "    return cluster_assignments, cluster_centers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "assignments, centers = calc_kmeans(100, features_scaled,)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "17207.945644503612"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Calculate Calinski-Harabasz score (larger is better)\n",
    "metrics.calinski_harabaz_score(features_scaled, assignments)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "732.31562691166084"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def sum_squares(n_clusters, cluster_assignments):\n",
    "    \"\"\"Calculates the sum of squared errors.\"\"\"\n",
    "    square_sums = np.zeros(n_clusters)\n",
    "    for i in range(n_clusters):\n",
    "        indexes = np.where(cluster_assignments == i)[0]\n",
    "        # Squared errors\n",
    "        s = np.square(np.subtract(features_scaled[indexes], centers[i]))\n",
    "        # Sum of squared errors\n",
    "        square_sums[i] = np.sum(s)\n",
    "    return np.average(square_sums)\n",
    "\n",
    "sum_squares(100, assignments)"
   ]
  }
 ],
 "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": "3.6.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}