The Essential k-NN Algorithm

Overview

We can summarize the $k$-NN algorithm as having the following steps:

1. Create a list of all (distance, training sample) pairs.
2. Sort these in ascending order.
3. Pick to the first $k$, which will be the k nearest neighbors.
4. Chose the mode (the highest frequency) label for the $k$ nearest neighbors.

The implementation would look like this: