摘要：Data mining provides biological research a useful information analyzing tool. The key factors which influence the performance of biological data mining approaches are the large-scale of biological data and the high similaritiesamong patterns mined. In this paper, we present an efficient algorithm named IRTM for mining frequent subtrees embedded in biological data. We also advance a string encoding method for representing the trees, and a scope-list for extending all substrings for frequency test. The IRTM algorithm adopts vertically mining approach, and uses some pruning techniques to further reduce the computational time and space cost. Experimental results show that IRTM algorithm can achieve significantly performance improvement over previous works.
关键词：Embedded Frequent Sub Tree;Scope -List;Biological data
