Compressing the Set of Frequent Sequential Patterns

Compressing the set of frequent sequential patterns is a method in order to address the problem of explosive number of output sequential patterns. In order to get high-quality compression, it first clusters frequent sequential patterns, and then select and output only a representative sequential pattern for each cluster such that the number of these representative sequential patterns is minimized. A greedy algorithm and an efficient candidate based algorithm are proposed. The set of representative sequential patterns is a kind of subset of frequent sequential patterns. Experimental results show that it can achieve very good compression effect.