The service recommender is a subcomponent of the service management infrastructure. The aim of this module is to assess which are the best services or service objects according to several criteria that matches users’ needs. By taking a service description as input, the recommendation is implemented through a four-step process composed of similarity evaluation, filtering, scoring and ranking activities. The recommender returns a ranked list of services, which are similar to the input one. Top-ranked services are the best according to recommendation criteria. For the final prototype of the recommender, filtering and scoring are implemented for evaluating the following criteria: (i) service trust, (ii) service usage and (iii) sensor status of service objects. The main new feature of the final prototype is a similarity engine, which provides support for content-based recommendation of services. In addition, criteria evaluation techniques have been improved in terms of both effectiveness and efficiency. This document describes the high level architecture of the recommender and the novel algorithms and techniques that implement the overall recommendation process for the final prototype.
D31.2.2 Service recommender – Final prototype