Cluster analysis with cellwise trimming and applications for the robust clustering of curves

Abstract

In this work, we propose a robust Cluster Analysis methodology based on cell trimming as an extension to a recently introduced robust version of Principal Component Analysis. This new approach allows for cellwise trimming in cluster analysis, which is more reasonable than traditional casewise trimming when the problem's dimension is large. This type of trimming avoids an unnecessary loss of information when only a few cells of the entirely trimmed observations are atypical. An algorithm is proposed to apply this approach. This algorithm is particularized to the interesting case of functional cluster analysis. Simulations and applications to real data sets are given to illustrate the proposed methods.