Biclustering, block clustering,[1][2] Co-clustering or two-mode clustering[3][4][5] is a data mining technique which allows simultaneous clustering of the rows and columns of a matrix. The term was first introduced by Boris Mirkin[6] to name a technique introduced many years earlier,[6] in 1972, by John A. Hartigan.[7]
Given a set of samples represented by an -dimensional feature vector, the entire dataset can be represented as rows in columns (i.e., an matrix). The Biclustering algorithm generates Biclusters. A Bicluster is a subset of rows which exhibit similar behavior across a subset of columns, or vice versa.