WS-Miner: A Fast Weighted Sequential Pattern Mining Algorithm

Traditional sequence pattern mining research focuses on mining frequent sequence patterns from sequence databases. However, the frequency of patterns alone does not determine their importance. That means frequent patterns may not be needed, while infrequent patterns may hold high importance. In this paper, an effective algorithm for weighted sequential pattern mining, called WS-Miner, is proposed. The novel upper bound can effectively reduce the search space during the mining process, and an optimized projection technology is utilized to address the memory explosion issue. Finally, extensive experiments are conducted to exhibits superior performance in terms of runtime and memory cost, which demonstrates the new algorithm can discover all interesting patterns effectively.