Kmeans-no-test.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "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": [
    "a = np.array ([0.1707, 0.2293])\n",
    "b = np.array ([0.2293, 0.2610])\n",
    "c = np.array ([0.2439, 0.1463])\n",
    "e = np.array ([0.7000, 0.2439])\n",
    "d = np.array ([0.6171, 0.2414])\n",
    "f = np.array ([0.6683, 0.2536])\n",
    "g = np.array ([0.7878, 0.8219])\n",
    "h = np.array ([0.8488, 0.8609])\n",
    "i = np.array ([0.8732, 0.8536])\n",
    "j = np.array ([0.6195, 0.2634])\n",
    "k = np.array ([0.3707, 0.5293])\n",
    "l = np.array ([0.4293, 0.7010])\n",
    "m = np.array ([0.4039, 0.6463])\n",
    "patterns = [a, b, c, d, e, f, g, h, i, j, k, l,m]\n",
    "names = ['a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i', 'j']\n",
    "for sample in patterns:\n",
    "        plt.plot(sample[0], sample[1],'bo', markersize = 8)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# returns a numpy array like:\n",
    "# [[0.6683 0.2536]\n",
    "#  [0.6195 0.2634]\n",
    "#  [0.7    0.2439]]\n",
    "\n",
    "def generate_centroids(X, cluster_count):\n",
    "    np.random.RandomState(123)\n",
    "    index = np.random.permutation(X.shape[0])\n",
    "    centroids = X[index[:cluster_count]]\n",
    "    return centroids"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "def find_closest_cluster(samples, centroids):\n",
    "    all_labels = [] # a label for each sample\n",
    "    samples = np.array(samples)\n",
    "    centroids = np.array(centroids)\n",
    "    for i in range(samples.shape[0]): # iterating over samples\n",
    "        sample = samples[i]\n",
    "        label = []\n",
    "        for j in range(centroids.shape[0]): # iterating over centroids\n",
    "            cent = centroids[j]\n",
    "            d = (sample[0]-cent[0])**2 + (sample[1]-cent[1])**2\n",
    "            label = np.append(label, d)\n",
    "        all_labels = np.append(all_labels, np.argmin(label))\n",
    "        \n",
    "    return all_labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "def find_centroids(samples, labels, cluster_count):\n",
    "    samples = np.array(samples)\n",
    "    for i in range(cluster_count):\n",
    "        cent = samples[labels == i].mean(axis = 0)\n",
    "        if i == 0:\n",
    "            centroids = [cent]\n",
    "        else:\n",
    "            centroids = centroids + [cent]\n",
    "            \n",
    "    return centroids"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "def train(cluster_count, iteration_count, samples):\n",
    "    \n",
    "    samples = np.array(samples)\n",
    "    centroids = generate_centroids(samples, cluster_count)\n",
    "    # plot initial situation\n",
    "    plt.subplot(1, 2, 1)\n",
    "    for sample in patterns:\n",
    "        plt.plot(sample[0], sample[1],'bo', markersize = 8)\n",
    "    for sample in centroids:\n",
    "        plt.plot(sample[0], sample[1],'r^')\n",
    "        \n",
    "    for i in range(iteration_count):\n",
    "        labels = find_closest_cluster(samples, centroids)\n",
    "        centroids = find_centroids(samples, labels, cluster_count)\n",
    "    \n",
    "    #plot after situation \n",
    "    plt.subplot(1, 2, 2)\n",
    "    for sample in patterns:\n",
    "        plt.plot(sample[0], sample[1],'bo', markersize = 8)\n",
    "    for sample in centroids:\n",
    "        plt.plot(sample[0], sample[1],'r^')\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def calculate_J(samples, centroids)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "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": [
    "train(4, 1000, patterns)"
   ]
  },
  {
   "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": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}