A collaborative filtering algorithm with clustering for personalized web service selection in business processes

Recommender systems aim to propose items that are expected to be of interest to the users. As one of the most successful approaches to building recommender systems, collaborative filtering exploits the known preferences of a group of users to formulate recommendations or predictions of the unknown preferences for other users. In many cases, collaborative filtering algorithms handle complex items, which are described using hierarchical tree structures containing rich semantic information. In order to make accurate recommendations on such items, the related algorithms must examine all aspects of the available semantic information. Thus, when collaborative filtering techniques are employed to adapt the execution of business processes, they must take into account the services' Quality of Service parameters, so as to generate recommendations tailored to the individual user needs. In this paper, we present a collaborative filtering-based algorithm which takes into account the web services' QoS parameters in order to tailor the execution of business processes to the preferences of users. An offline clustering technique is also introduced for supporting the efficient and scalable execution of proposed algorithm under the presence of large repositories of sparse data.