Web service ranking based on context

The methods for document or web page search can not be adapted well for the efficient service search for the lacking of service descriptions. Importing new information related to services to help service search is an urgent task, which can support the development of SOA (service oriented architecture). One of the popularly used information is service context which describes service-related usages. It has been verified that context information is useful for improving the performance of search. However, previous research work focuses on using context as query content or result filters. In this work, we contribute to design a novel service ranking algorithm based on context information and propose to order services by both the content-based similarity and the context-based usefulness. In other words, we expect to return users the term-level relevant services which are the most useful ones. We first model the context by the weighted bipartite graphs, representing the relationship between services and the involved applications. Relying on the graphs, we design an iterative algorithm for evaluating service usefulness and then we combine it with content-based relevance value to get the final ranking values. Extensive experiments compare the recent work with ours on service search, and show our superior performance.