K Nearest Neighbours

Theory

Overview

In this blog, I will be explaining the K Nearest Neighbours algorithm. This is a classification algorithm that can predict which class a new point belongs to.

The way it works is by finding the Euclidean distance from the new point to each other points in the dataset. Then after taking the k closest points, you need to rank all of the labels of those points. The label that is most frequent in those k points is the one that is predicted for the new point.