Skip to content

Latest commit

 

History

History
11 lines (6 loc) · 929 Bytes

README.md

File metadata and controls

11 lines (6 loc) · 929 Bytes

K-means

This repository covers the implementation of K-Means algorithm for the assignment of AI course at UFAZ.

Requires: Iris (flower) dataset

The main idea is to classify data that are not necessary labelled. K-means will partition the dataset in k groups (or clusters).

When you have a labelled dataset, like in this case, you know how many clusters you would want to set in your program. When you do not have labelled data, you still have to propose a value for k: such models are referred to as semi-supervised models. Obviously, different values for k will lead to different results. Without labels, you don’t know a priority which value to use (k can be approximated using dedicated methods, click here to see examples on this subject).

For further information about the practical work read the PDF file describing all activities.