High Utility Partial Periodic Pattern Mining
Tzung‐Pei Hong, Jen-Hao Hsu, Guo-Cheng Lan, Kung-Jiuan Yang, Shyue-Liang Wang, Jerry Chun‐Wei Lin · 2017
In this paper, we study the properties of partial periodic pattern mining and extend the original problem to high-utility partial periodic pattern mining (HUPPP), which considers not only the occurring time order and periodic length of events but also the quantities and individual profits of the events. Based on the periodic utility function, we have presented a mining algorithm for finding high-utility partial periodic patterns. The algorithm uses the two-phased periodic utility upper-bound (PUUB) model to avoid information loss in the mining process. Finally, the experiments made to show the performance of the proposed algorithm under various parameter settings.