Projection pursuit density estimation and cluster analysis: an interaction

Abstract

Since it does not require large sample size to dimensionality ratios, projection pur-suit density estimation (ppde) could enlarge the applicability of clustering methods based on nonparametric density estimation. In this paper we attempt to evaluate and compare the effectiveness of ppde when used in Wong and Lane's clustering procedure (1983) instead of the nearest neigbour density estimator. In order to adapt their algorithm to ppde, a new clustering criterion has been proposed; its performances have been investigated through a simulation study and illustrated on a real data set.