ABSTRACT
Cheng and Church (2000) introduced biclustering as simultaneous clustering of both features and samples in order to find underlying patterns. The patterns are characterised by biclusters with mean squared residual score smaller than a pre-specified threshold (δ). Yang et al. (2003) proposed FLOC (Flexible Overlapping biClustering), an alternative algorithm to find δ-biclusters without the interference of random data used to mask a found bicluster or missing values in the approach of Cheng and Church (2000).
