Clustering for n00bies

Machine Leaning textbooks can leave someone that isn’t used to hearing about eigenvectors more than a little bit put-off. Although the majority of the algorithms are designed to perform tasks we’re all intimately familiar with (like classification and clustering), somehow, the intuition behind the techniques gets lost.

Also, instead of examples that a reader can relate to, greek letters, fancy symbols and letters like x,y,i,j,k,n, and m command the page. Some books have been more successful in the explanations, they fall short with supplementary code that isn’t Matlab.

In his post, k-Nearest Neighbor for Humans, James Robert does a great job of doing the impossible — explaining clustering methods in simple terms, and providing the detailed steps to repeat the analysis. Replicating his steps with his data at first, and then yours, is a great way to get your feel wet.

Happy clustering!