ONE TIME ENUMERATION OF MAXIMAL BICLIQUE PATTERNS FROM 3D SYMMETRIC MATRIX

摘要

We propose an algorithm Cubeminer-MBC*, to extract maximal biclique patterns from a 3D symmetric adjacency matrix only once. In this paper, we introduce (i) a novel enumeration strategy and (ii) a new pruning strategy, which results 50% reduction in search space and maximal biclique patterns are generated only once, i.e., zero duplicates are generated. On the basis of experiments conducted, we observed Cubeminer-MBC* outperforms Cubeminer in terms of running time.