Data Providing Web Service Selection Using Bayesian Network

Web services can be broadly classified into two types, namely effect providing (EP) services and data providing (DP) services. Often times, several DP services need to be composed to provide consolidated data from different sources, which in turn serves as the input for some EP service. There have been quite some works that consider data semantics in composing DP services and assure that the composed DP services are compatible with each other in their data semantics. When there are several composition plans, it is not clear which one is the best. In this work, we address the DP service composition problem that intends to satisfy user preference specified at the instance level, namely the expected occurrence. We first use the query rewriting method to identify a composition of service types that satisfies user´s requirement and employ Bayesian network model to express the causal relationship between exchange variables of DP service types. Service selection is then conducted by computing the posterior probability in the Bayesian network. We conduct experiments to show how our proposed Bayesian network-based method performs with respect to the other baseline methods, in terms of execution success rate, data quality, and execution time.