A FOAF-based framework for e-commerce recommender service system

Recommender service systems have been widely and successfully applied in e-commerce to provide personalized recommendations to customers nowadays. The tremendous growth in the amount of available information and the number of visitors to websites poses some challenges for recommender service systems such as poor prediction accuracy, scalability, and dynamic changes of users. To address these issues and increase the performance of the systems, an e-commerce recommender service system framework based on FOAF (Friend of A Friend) is proposed in this paper. FOAF provides a RDF/XML vocabulary to describe individual users and their relationships with other users. A FOAF profile could allow a system to better understand users' personalized needs. Once the system extracts each user's preferences that are represented by FOAF document format, it can classify users with respect to their own preferences on real time, and also, it can recommend items in which some users are interested to a target user who has the highest similarity with them in the same group. Experimental results show that the proposed framework helps to reduce the recommendation time, while improving accuracy.