A personalized approach to experience-aware service ranking and selection

Abstract

Existing approaches to service ranking and selection evaluate the suitability of available services for a given request based on the advertisement created by the service provider. They will compare how well the advertisement matches the service request and will choose the service with the best matching advertisement. Unfortunately, at this point in time, it is uncertain whether the service that will actually be performed will match the request as well as the advertisement promised. In this paper, we present an approach that reduces the degree of this uncertainty by taking previous experiences with the service provider (which reflect the performance of the actual service not the advertisement) into account. Contrary to many other approaches our solution accounts for the subjective nature of rating-based experiences by considering the preferences of the experience creators. Moreover it exploits the number of available experiences more effectively by considering not only experiences for a given service, but also experiences for similar services of the same provider. Our solution utilizes indirect user information and avoids explicit sharing of personal consumer information. © 2008 Springer-Verlag.