A Kth Nearest Neighbour Clustering Procedure (Classic Reprint) - Softcover

Inhaltsangabe

This book presents a novel clustering procedure applicable to both variable-by-case data matrices and case-by-case dissimilarity matrices. Unlike many existing methods that approach clustering as a heuristic, the proposed procedure is based on a uniformly consistent density estimate, making it useful for drawing statistical inferences about the underlying population from a sample. The author demonstrates the asymptotic consistency of the method for high-density clusters in several dimensions and illustrates its small-sample behavior through empirical examples. A real-world application showcases the practical utility of this method, which provides valuable insights into identifying high-density clusters in data.

Die Inhaltsangabe kann sich auf eine andere Ausgabe dieses Titels beziehen.