K-means

This project is the implementation of K-means clustering algorithm for unsupervised machine learning.

Task from Machine Learning for Greenhorns – Winter 2020/21. Part of this code was provided by Milan Straka.

k_means

The algorithm is implemented in k_means.py. The main method is k_means.

k_means_main

The file k_means_main.py shows how to use the implemented algorithm on some artificial data.

The example of the invocation of the program is:

python k_means_main.py --clusters=7 --examples=200 --iterations=5 --seed=67 --init=kmeans++ --plot=plot

• clusters specifies the number of clusters to generate
• examples specifies the amount of data points to be generated
• iterations is the number of iterations of the K-means algorithm
• seed specifies the random seed
• init specifies the initialization of the cluster centers, can be random or kmeans++
• plot is the name of output file of the plot

The result of the previous invocation is the following output file: