Author(s)

E. Maltseva, C. Pizzuti & D. Talia

Abstract

Clustering is an undirected knowledge discovery technique based on the partitioning of large sets of data objects into homogenous groups. All ob- jects contained in the same group have similar characteristics. Grouping multivariate data is a difficult data mining task when no domain knowl- edge on data structure is available. In this paper we describe the use of a well known linear projection technique, called singular value decomposition (SVD), to discover clusters in the pattern space by projecting it into a sub- space that constitutes its best approximation and preserves the character of data. Experimental results on real datasets from the UCI Machine Learning repository assess the quality of the clustering obtained. 1 Introduction Clustering is a data mining task [1] for unsupervised classification that consists in p

Keywords