Lecture
This lecture introduces the k-Nearest-Neighbor (kNN) classifier, which classifies a point based on the majority label of its k nearest neighbors in the training set. It covers the improved algorithm, hand-written digit recognition, validation, multi-class k-NN, data reduction, condensed nearest neighbors, and applications like recommender systems. The lecture also discusses the construction of kNN graphs, greedy algorithms, gossip-based computing, limitations of kNN, distribution of distances in 2D and N-D, volume of a sphere in N-D, and the curse of dimensionality in machine learning.