K-Nearest Neighbours - 'KNN', is a supervised Learning algorithm. It uses proximity to make classifications.
Hit '>Play' above to learn more about how KNN helps classify multi-dimensional data.
Points in a multi-dimensional feature space are assigned to two or more categories. Unclassified points to be assigned to one of these categories are placed in this same feature space.
A hyper-parameter 'K' is provided. The distances are calculated between our unclassified point and its 'K' Nearest Neighbours. The unclassified point will be assigned to the majority (if binary) or plurality (if three of more) category of these nearest neighbours. KNN is one of the earliest - and simple - Machine Learning Techniques.
Developed in 1951 it adopts the 'birds of a feather' principle. KNN is extensively used in multiple industries; including FinServ, healthcare, eCommerce as well as in psephology and political campaigns. KNN algorithms are used to recommend products on Amazon, movie & tv content on Netflix and for online ad display.