K-Means Clustering

Platform

Ubuntu 16.04

Library Used:

sklearn

pip install sklearn

numpy

pip install numpy

Running Program

I have created two experiments

To run first experiment:

python main.py 1 <NUMBER OF CLUSTERS> <TO_PRINT>   
e.g main.py 1 10 FALSE

where:
<NUMBER OF CLUSTERS> is the number of clusters
<TO_PRINT> is a BOOLEAN VALUE TRUE/FALSE whether to print the assigned label and actual label for each instance
Note Setting <TO_PRINT> to TRUE would print 5000 lines of assigned and actual label for each instance

Output is Accuracy and The mapping between random label given by k means and actual assigned label   
where actual assigned label is the max frequency label in that random label.
See Report for more information

To run Second experiment:

python main.py 2

This would run the k-means for varying number of clusters in k=5 to 100. (You may Change this in Line 27 of main.py if you wish to)

This would print the accuracy for each varying cluster and show a graph as well