Effective Pruning Strategies for Sequential Pattern Mining

In this paper, we systematically explore the search space of frequent sequence mining and present two novel pruning strategies, SEP (Sequence Extension Pruning) and IEP (Item Extension Pruning), which can be used in all Apriori-like sequence mining algorithms or lattice-theoretic approaches. With a little more memory overhead, proposed pruning strategies can prune invalidated search space and decrease the total cost of frequency counting effectively. For effectiveness testing reason, we optimize SPAM [2] and present the improved algorithm, SPAM
_SEPIEP
, which uses SEP and IEP to prune the search space by sharing the frequent 2- sequences lists. A set of comprehensive performance experiments study shows that SPAM
_SEPIEP
outperforms SPAM by a factor of 10 on small datasets and better than 30% to 50% on reasonably large dataset.