Even Faster Exact k-Means Clustering

Abstract

A naïve implementation of k-means clustering requires computing for each of the n data points the distance to each of the k cluster centers, which can result in fairly slow execution. However, by storing distance information obtained by earlier computations as well as information about distances between cluster centers, the triangle inequality can be exploited in different ways to reduce the number of needed distance computations, e.g. [3–5, 7, 11]. In this paper I present an improvement of the Exponion method [11] that generally accelerates the computations. Furthermore, by evaluating several methods on a fairly wide range of artificial data sets, I derive a kind of map, for which data set parameters which method (often) yields the lowest execution times.