K-Means-Code

Implementation of K Means Algorithm in Python

Getting Started

Pre Requisites

pandas
numpy
matplotlib

Usage

1. Using main.py

Takes 3 arguements length of sample data number_of_datapoints , number of clusters number_of_clusters and number of iteration to performnumber_of_iterations

python main.py number_of_datapoints number_of_clusters number_of_iterations  

2. Using simplekmeans.py

File Contains kmeans function def KMeans(data,k,iters) and sample usage

data=np.random.rand(100,2)
labels,centroids=KMeans(data,4,20)

File Guide

• Folder: code-at-1-hour-mark -> contains the code at 1 hour mark
• File: simplekmeans.py -> contains function that fits data on specified clusters and iterations
• File: main.py -> use this file to run program , specify the number of datapoints to randomly generate, number of clusters and number of iterations
• File: algorithm/mlearn.py -> contains KMeans class used by main.py
• Folder: output -> contains output of main.py ie plots,data files

Algorithm

1. Generate data points
2. Calculate kmeans
   1. select k random points from data
   2. repeat until iterations
      1. initialize centroids with points selected in above step
      2. assigning data points to centroids
      3. re initialize centroids with mean of selected data points
   3. return labels,centroids
3. Make corresponding plots and save data to output dir

Plots

Initial Data

Label generated against specified iterations and clusters