K-Means-Cluster-Analysis

A simple program which performs K-Means clustering on a data set,visualizes the results and calculates validity metrics.

Usage

• Set number of max iterations(max_iter) in the file main.m.
• Save the data in the file "data.mat". Data must be a N x 2 matrix where each row contains X and Y coordinate
• Run main.m and enter 'K' value to begin the clustering. K must be a positive integer less than or equal to the number of data points.
• The program will plot the results as well as display the values of the validity metrics.
  Note: To generate a new random data set of 500 points, run the GetNewData.m file.

Features

• Plots Initial Data.
• Calculates optimal K-Mean centres corresponding to local minimum and plots the location.
• Plots each cluster with a characteristic random colour.
• Calculates the following validity metrics:
  • Compactness Value
  • Separation Value
  • Davies Bouldin Index
  • Dunn Validity Index