High Fuzzy Utility Based Frequent Patterns Mining Approach for Mobile Web Services Sequences

Abstract

Nowadays high fuzzy utility based pattern mining is an emerging topic in data mining. It refers to discover all patterns having a high utility meeting a user-specified minimum high utility threshold. It comprises extracting patterns which are highly accessed in mobile web service sequences. Different from the traditional fuzzy approach, high fuzzy utility mining considers not only counts of mobile web service accessed in a sequence but additionally their preference value while mobile web services sequences are accessed. In this paper, I introduce a new approach, namely HFUBPM (High Fuzzy Utility Based Patterns Mining) for high fuzzy utility patterns extraction from mobile web services accessed sequences. The proposed approach uses a fuzzy minimum operator to extract highly interesting patterns from web service accessed sequences. In this proposed approach, downward closure property in fuzzy sets is handled by an efficient upper bound model. This model improves the efficiency of mining way. At last, the experiments have been performed on both synthetic and real datasets, which show that the proposed approach has good performances in terms of execution and search space.