The negative utility values of items are not considered and the efficiency of streaming data processing in high utility pattern mining algorithm is not acceptable.An efficient high utility pattern mining algorithm with negative unit profit based on sliding window(HUPN_SW)was proposed.A novel utility list structure was designed,namely positive and negative with sli-ding window utility list(PNSWU-List).The high utility itemsets could be efficiently mined from the latest data in the sliding window without multiple database scans and candidate generation.HUPN_SW was effective for dynamic streaming data.Experi-mental results show that HUPN_SW algorithm exhibits good performance in terms of running time and scalability.
维普
万方数据