Composition Context Matching for Web Service Recommendation

In a service oriented B2B context, companies are aiming to cooperate and interact with others to dynamically create added value services by composing them. The tremendous growth in the amount of available web services impulses many researches on proposing discovery mechanisms to help designers compose services. However, most of the proposed solutions ignore in their discovery queries the upcoming services' interactions within the being designed compositions. In this paper, we present an original recommendation approach to discover a requested service to be incorporated into a specific composition. We propose to take into account the composition context specified through the process fragment surrounding the requested service, and benefit from the modeling and usage of previous service compositions to build our recommendations. Provided experimental evaluations in this paper show that our approach is efficient in realistic situations.