Abstract:

Since there exist redundant sequential patterns in results, such mining runs for a long time. To combat this drawback, a new algorithm, called closed sequential patterns mining algorithm in time order (CloTSP),is proposed. Based on the nature of closed sequential patterns, CloTSP can judge whether a frequent sequential pattern is extended by comparing its time order with each frequent 1-item subsequence. Experiments on the synthetic data produced by International business machines corporation sequential pattern generator show that CloTSP can shorten run-time significantly compared to closed sequential pattern mining (CloSpan). Furthermore, it is also shown that the run-time of CloTSP is not affected by variation of attribute numbers.

Key words: time order, closed sequential patterns, data mining