A Mining Algorithm for Fast Maximal Sequential Patterns

This paper proposes a novel algorithm MFSPAN(maximal frequent sequential pattern mining(algorithm)).MFSPAN is used to mine the complete set of maximal frequent sequential patterns in sequence(databases).It solves the problem that the number of frequent subsequences will increase explosively as frequent(patterns) become longer: because MFSPAN takes full advantage of the property that different sequences may share a common prefix to reduce itemset comparing times.Experiments on standard test data show that(MFSPAN) is very effective.