Recommender systems, as effective approaches to information overload issue, have attracted researchers’ attention especially in the field of electronic commerce. They provide personalized recommendations on products/services to customers. From a seller’s point of view the process of recommendation is a multi-criteria decision problem. So the recommender systems can also be seen as decision support systems which help a seller to decide and choose suitable items for recommendation. This study proposes a novel design of recommender systems for product recommendation to new users -by collecting both implicit and explicit information- based on the PROMETHEE II methodology which is a well known multi-criteria decision analysis out-ranking method. We also experienced the recommendation from different but relevant sorts of items in this design. The performance of the proposed recommendation algorithm is tested with real world data. Experimental results reveal that the proposed system is feasible and can yield satisfactory recommendations.
展开▼
