KMeansPlusPlus

Python K-means++ implementation

About

This implementation is more for my own purposes and to get a better idea of how K-Means++/K-Means work

Requirements

• Python 3.x
• Numpy
• Matplotlib (optional, to run Testing.py and graph results)

Usage

The constructor takes in two arguments, points_list and k:

• points_list: 2d list containing n-dimensional points

  points_list = [[Point1], [Point2], etc.]; n-dimensional point x = [X1, X2, ..., Xn]

• k: Number of desired clusters/centroids

Calling final_centroids after creating the object returns the final centroids in the same format as points_list

Example

Seeds from randomly generated data, x∈[0, 100], y∈[0, 50], n = 50, k = 7:

Notes

Testing.py is used to create the screenshots, but is not necessary to run K-Means++