K-means clustering.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "55923492",
   "metadata": {},
   "outputs": [],
   "source": [
    "class kmeans:\n",
    "\n",
    "  def __init__(self, k, max_iter=None):\n",
    "    self.k = k # stores the number of the clusters.\n",
    "    self.max_iter = max_iter # stores the maximun iteration allowed for the task.\n",
    "    self.closest_clusters = None # stores the closest cluster (from 0 to 39) for every point.\n",
    "    self.cluster_centers = None # stores all the cluster centers.\n",
    "\n",
    "  def fit(self, x_train, centers=[]): #initialising the centeres are an empty list if there was no input for initial cluster centers.\n",
    "\n",
    "    x_train = np.array(x_train)\n",
    "\n",
    "    if len(centers) == 0: # checks if the initial cluster centers were not given as input.\n",
    "      centers = x_train[np.random.randint(low=0, high=x_train.shape[0], size=self.k)] # set the centers as a k size array which has centers.\n",
    "\n",
    "    if self.max_iter == None: # if there is an upper bound for maximun number of iterations.\n",
    "      while 1: # the loop will continue until there is a break statement.\n",
    "        distances = distance.cdist(x_train, centers) # to calculate distance from every point in \"x_train\" to every point in \"centers\" and store in an array named \"distances\".\n",
    "\n",
    "        closest_cluster = [] # stores the closest clusters indices for every point.\n",
    "        for i in distances:\n",
    "          closest_cluster.append(i.argmin()) # adds the index closest cluster to the list among all the clusters.\n",
    "        closest_cluster = np.array(closest_cluster)\n",
    "\n",
    "        new_centers = centers.copy() # new_centers is a copy of centers.\n",
    "        for i in range(self.k): # loop runs for all the clusters.\n",
    "          if i in closest_cluster: # checks if a point has that cluster as the closest cluster.\n",
    "            new_centers[i] = np.mean(x_train[closest_cluster == i], axis=0) # the center is changed to mean of the points in the cluster corresponding to that center.\n",
    "        new_centers = np.array(new_centers)\n",
    "\n",
    "        if np.array_equal(new_centers, centers): # checks if the initial centers and the new centers are same.\n",
    "          break # if the centers does not change, breaks the loop.\n",
    "\n",
    "        centers = new_centers # centers are changesd to new_centers.\n",
    "\n",
    "    else: # if there is an upper bound for iterations.\n",
    "      for i in range(self.max_iter): # loop runs for that max_iter number of iterations.\n",
    "        distances = distance.cdist(x_train, centers) # to calculate distance from every point in \"x_train\" to every point in \"centers\" and store in an array named \"distances\".\n",
    "\n",
    "        closest_cluster = [] # stores the closest clusters indices for every point.\n",
    "        for i in distances:\n",
    "          closest_cluster.append(i.argmin()) # adds the index closest cluster to the list among all the clusters.\n",
    "        closest_cluster = np.array(closest_cluster)\n",
    "\n",
    "        new_centers = centers.copy() # new_centers is a copy of centers.\n",
    "        for i in range(self.k): # loop runs for all the clusters.\n",
    "          if i in closest_cluster: # checks if a point has that cluster as the closest cluster.\n",
    "            new_centers[i] = np.mean(x_train[closest_cluster == i], axis=0) # the center is changed to mean of the points in the cluster corresponding to that center.\n",
    "        new_centers = np.array(new_centers)\n",
    "\n",
    "        if np.array_equal(new_centers, centers) == True: # checks if the initial centers and the new centers are same.\n",
    "          break # if the centers does not change, breaks the loop.\n",
    "\n",
    "        centers = new_centers # centers are changesd to new_centers.\n",
    "    \n",
    "    self.closest_clusters = closest_cluster\n",
    "    self.cluster_centers = centers"
   ]
  }
 ],
 "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": "3.11.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}