Constructing Service Semantic Link Network Based on the Probabilistic Graphical Model

Abstract

Automatic services collaboration calls for the development of semantically structured service network to maximize the utility of Web services. Service Semantic Link Network (S-SLN) is the semantic model for effectively managing Web service resources by the dependency relationship between services. We provided an effective method for constructing S-SLN based on the graphical structure representation of the dependencies embedded in a probabilistic model. A Markov network is an undirected graph whose links represents probability dependencies. We first learned Markov network structure from Web services data, and then transformed the undirected Markov network structure into a directed graph structure of S-SLN based on the same joint probability distribution. Finally, experimental results show the effectiveness of the method.