Belief revision for intelligent web service recommendation

With the increasing number of web services deployed to the world wide web these days, discovering, recommending, and invoking web services to fulfil the specific functional and preferential requirements of a service user has become a very complex and time consuming activity. Accordingly, there is a pressing need to develop intelligent web service discovery and recommendation mechanisms to improve the efficiency and effectiveness of service-oriented systems. The growing interests in semantic web services has highlighted the advantages of applying formal knowledge representation and reasoning models to raise the level of autonomy and intelligence in human-to-machine and machine-to-machine interactions. Although classical logics such as description logic underpinning the development of OWL has been explored for services discovery, services choreography, services enactment, and services contracting, the non-monotonicity in web service discovery and recommendation is rarely examined. The main contribution of this paper is the development of a belief revision logic based service recommendation agent to address the non-monotonicity issue of service recommendation. Our initial experiment based on real-world web service recommendation scenarios reveals that the proposed logical model for service recommendation agent is effective. To the best of our knowledge, the research presented in this paper represents the first successful attempt of applying belief revision logic to build adaptive service recommendation agents. © 2012 Springer-Verlag Berlin Heidelberg.