Query-based Biclustering using Formal Concept Analysis.

摘要: Biclustering methods have proven to be critical tools in the exploratory analysis of high-dimensional data including information networks, microarray experiments, and bag words data. However, most biclustering fail answer specific questions interest do not incorporate prior knowledge expertise from user. To this end, query-based algorithms that are recently developed context utilize a set seed genes provided by user which assumed tightly co-expressed or functionally related prune search space guide algorithm. In paper, novel QueryBased Bi-Clustering algorithm, QBBC, is proposed new formulation combines advantages low-variance techniques Formal Concept Analysis. We prove statistical dispersion measures order-preserving induce an ordering on biclusters turn, exploited form efficient manner. Our approach provides mechanism generalize sparse such as networks words. Moreover, framework performs local opposed global approaches previous employed. Experimental results indicate often produces higher quality precise compared state-of-the-art querybased methods. addition, our performance evaluation illustrate efficiency scalability QBBC full other existing approaches.