Pleiades: Subspace clustering and evaluation

Abstract

Subspace clustering mines the clusters present in locally relevant subsets of the attributes. In the literature, several approaches have been suggested along with different measures for quality assessment. Pleiades provides the means for easy comparison and evaluation of different subspace clustering approaches, along with several quality measures specific for subspace clustering as well as extensibility to further application areas and algorithms. It extends the popular WEKA mining tools, allowing for contrasting results with existing algorithms and data sets. © 2008 Springer-Verlag Berlin Heidelberg.