Machine Learning with k-Nearest Neighbors

K-nearest neighbors, while perhaps the simplest possible machine learning algorithm, is extremely useful for a variety of tasks. In this video, I explain how it works and demonstrate its performance on handwriting recognition.

See the handwriting demo here: http://macheads101.com/demos/handwriting/?c=neighbors

Source of handwriting demo: https://github.com/unixpickle/mnistdemo

MNIST handwriting dataset: http://yann.lecun.com/exdb/mnist/

Code to generate graphics: https://gist.github.com/unixpickle/9ff10601ab91e09afa04344b80b36c4e