Learning relaxed 3-clusters from pairs of related datasets

摘要: In many emerging data mining situations we encounter multiple large binary relational datasets that are generated independently but semantically interconnected and must be mined simultaneously to obtain an integrated effect of the residing in all them. The idea finding 3-clusters is increasingly used where one has concurrently mine two distinct share a common domain along dimension. By discovering 3-clusters, can important insights on underlying connections between objects different domains. All 3-clustering algorithms for presented till now able find bi-clusters 3-cluster strict, is, rectangle formed by contains only ‘1’ entries. However, real world applications very sparse relaxed bi-cluster, allows some zeros bi-clusters' rectangles, valuable. this paper, present novel search based algorithm finds from domain. Each identified involves whose overlap sets maximal. Through our algorithm, also exert finer control over percentage 1 s allowed each bi-clusters. We validate effectiveness using synthetic Our results show notion produce more meaningful when compared with strict requirement ones.