Proximity measures and results validation in biclustering - A survey

Abstract

The concept of biclustering evolved from traditional clustering techniques, which have proved to be inadequate for discovering local patterns in gene microarrays, in particular with shifting and scaling patterns. In this work we compare similarity measures applied in different biclustering algorithms and review validation methodologies described in literature. To our best knowledge, this is the first in-depth comparative analysis of proximity measures and validation techniques for biclustering. Current trends in design of similarity measures as well as a rich collection of state-of-the-art benchmark datasets are presented, supporting algorithm designers in classification of comparison and quality assessment criteria of emerging biclustering algorithms. © 2013 Springer-Verlag.