TKUS: Mining top-k high utility sequential patterns

High-utility sequential pattern mining (HUSPM) has recently emerged as a focus of intense research interest. The main task HUSPM is to find all subsequences, within quantitative database, that have high utility with respect user-defined minimum threshold. However, it difficult specify the threshold, especially when database features, which are invisible in most cases, not understood. To handle this problem, top-k was proposed. Up now, only very preliminary work been conducted capture HUSPs, and existing strategies require improvement terms running time, memory consumption, unpromising candidate filtering, scalability. Moreover, no systematic problem statement defined. In paper, we formulate propose novel algorithm called TKUS. improve efficiency, TKUS adopts projection local search mechanism employs several schemes, including Sequence Utility Raising, Terminate Descendants Early, Eliminate Unpromising Items strategies, allow greatly reduce space. Finally, experimental results demonstrate can achieve sufficiently good performance compared state-of-the-art TKHUS-Span.