A Recommendation Method of BPEL Activity Based on Association Rules Mining

Abstract

As the de facto standard, Business Process Execution Language (BPEL) is widely used in the composition and orchestration of web services; however, it is tiring and error-prone if we just rely on the developers to create every activity icon and assign it an atomic service. On the basis of the author's early work, this paper creatively applies association rules mining to the analysis of BPEL process, and put forwards a recommendation method of BPEL activity based on Labeled Activity Tree (LAT) and sub-activity sequence (subASeq). Finally, the experiment result of the BPEL processes in a multimedia conference system indicates this method's validity.