摘要：With the recent advanced computing, networking technologies and embedded systems, the computing paradigm has switched from mainframe and desktop computing to ubiquitous computing, one of whose visions is to provide intelligent, personalized and comprehensive services to users. As a new paradigm, Active Services is proposed to generate such services by retrieving, adapting, and composing of existing service components to satisfy user requirements. As the popularity of this paradigm and hence the number of service components increases, how to efficiently retrieve components to maximally meet user requirements has become a fundamental and significant problem. However, traditional facet-based retrieval methods only simply list out all the results without any kind of ranking and do not lay any emphasis on the differences of importance on each facet value in user requirements, which makes it hard for user to quickly select suitable components from the resulting list. To solve the problems, this paper proposes a novel personalized facet-weight based ranking method for service component retrieval, which assigns a weight for each facet to distinguish the importance of the facets, and constructs a personalized model to automatically calculate facet-weights for users according to their histo -rical retrieval records of the facet values and the weight setting. We optimize the parameters of the personalized model, evaluate the performance of the proposed retrieval method, and compare with the traditional facet-based matching methods. The experimental results show promising results in terms of retrieval accuracy and execution time.
