An efficient algorithm for mining maximal frequent sequences by Top-Down Delay Decomposition method

作者：Wang Qian; Ren Jiadong^*; Liu Qiang

来源：ICIC Express Letters, 2014, 8(7): 2007-2012.

• Bottom up methods
• Candidate sequences
• Database scans
• Delay decompositions
• Frequent sequences
• Maximal frequent sequences
• Research issues
• Topdown

Maximal frequent sequence mining is an important research issue which has realized the highly compressed storage of frequent sequences. At present, most algorithms are based on bottom-up method and large numbers of candidate sequences are generated. In this paper, an algorithm based on Top-down Delay Decomposition for mining Maximal Frequent Sequence is proposed. TDD MFS uses top-down delay decomposition method, sequences from original database and decomposition are expressed by different tuples to record their direct and indirect occurrence counts. Longest infrequent sequences are recursively decomposed and the decomposition of infrequent sequences that have frequent super sequences is avoided. Only one database scan is needed, no candidate sequences are generated and sequence decomposition times are reduced. Experiment result shows its high efficiency and good scalability.

• 出版日期2014