Lectures on the nearest neighbor method (Q892829)

The nearest neighbor paradigm is one of the most used machine learning methods in the classification domain. This book deals with different aspects regarding this approach, starting with the standard \(k\)-nearest neighbor model, and passing through the weighted \(k\)-nearest neighbor model, estimations for entropy, regression functions etc. Problems like point-wise and uniform consistency, rates of convergence, noiseless case etc. are also approached. The last part of the book deals with the supervised classification via \(k\)-nearest neighbors for fixed or variable parameter \(k\). The book ends with an appendix. It is intended for a large audience, including students, teachers, and researchers.