Efficient High Utility Negative Sequential Patterns Mining in Smart Campus

Abstract

High utility sequential patterns (HUSP) mining has been playing more and more important role in many applications, such as data analysis and smart campus. Current HUSP mining algorithms, however, only consider positive sequential patterns (PSP), do not consider negative sequential patterns (NSP). NSP mining, which takes non-occurring and occurring event into consideration, can play more important role than PSP in many applications. But current NSP mining algorithms haven’t considered utility. So in this paper, we propose a novel algorithm named HUNSPM to mine high utility negative sequential patterns (HUNSP). HUNSPM solves the key problems of how to calculate the utility of negative sequences, how to efficiently generate high utility negative sequential candidates (HUNSC) and how to store the HUNSC’s information. Substantial experiments show that HUNSPM can mine more HUNSP and use less time. To the best of our knowledge, HUNSPM is the first study that can mine HUNSP.