May 2019
Intermediate to advanced
162 pages
4h 24m
English
KNN is a really simple intuitive approach to building a clustering classifier. The idea is, given a set of labeled samples, when a new sample is introduced, we look at the k nearest points, and we make an estimate for its class membership based on the majority of points around it. So, in the following diagram we would classify this new question mark as a positive sample since the majority of its neighbors are positive:

Read now
Unlock full access