1. CoCluster: Efficient Mining Maximal Trend Biclusters Without Candidate Maintenance in Discrete Resource Effectiveness Matrix

Studying the level of efficiency of resources is a footstone for building the prognostics and health management system. This paper proposed an efficient bicluster mining algorithm—CoCluster, which mines trend bicluster in discrete resource effectiveness matrices. To improve the mining efficiency, it mines maximal trend bicluster using sample-growth method and multiple pruning strategies without candidate maintenance. Meanwhile, CoCluster algorithm can not only mine resource patterns with effectiveness in the downtrend, but also mine those with effectiveness in the uptrend, thus providing further decision support for later decision support system. To improve the generality, CoCluster algorithm can also mine resource patterns without change in effectiveness. The experimental results show our algorithm is more efficient than traditional algorithm.