TUSQ: Targeted High-Utility Sequence Querying

Significant efforts have been expended in the research and development of a database management system (DBMS) that has wide range applications for managing an enormous collection multisource, heterogeneous, complex, or growing data. Besides primary function (i.e., create, delete, update), practical impeccable DBMS can interact with users through information selection, is, querying their targets. Previous algorithms, such as frequent itemset sequential pattern (SPQ) focused on measurement frequency, which does not involve concept utility, is helpful to discover more informative patterns. To apply technology wider applications, we incorporate utility into target-oriented SPQ formulate task targeted utility-oriented sequence querying. address proposed problem, develop novel algorithm, namely high-utility (TUSQ), based two upper bounds (suffix remain terminated descendants utility) well vertical last instance table. For further efficiency, TUSQ relies projection utilizing compact data structure called chain. An extensive experimental study conducted several real synthetic datasets shows algorithm outperformed designed baseline terms runtime, memory consumption, candidate filtering.