ISBN-13: 9783639150995 / Angielski / Miękka / 2009 / 200 str.
Provision of personalized recommendations to usersrequires accurate modeling of their interests andneeds. This book presents a general framework andspecific methodologies for enhancing the accuracy ofuser modeling in recommender systems by importing andintegrating data collected by other recommendersystems. Such a process is defined as the mediationof user models and user modeling data. This bookdiscusses the details of the generic user modelingmediation framework, provides a user modeling datarepresentation model, demonstrates the compatibilityof the model with existing recommendation techniques,and discusses the general steps of the mediation.Then, the mediation framework is applied andillustrated with two practical mediation scenarios:cross-technique mediation and cross-domainmediation. Empirical evaluation of these scenariosshows that the mediation of user modeling data ispractical and beneficial, as it allows upgrading thequality of the recommendations provided to the users.