Using Personality Information in Collaborative Filtering for New Users

Recommender systems help users more easily and quickly find products that they truly prefer amidst the enormous volume of information available to them. Collaborative filtering (CF) methods, making recommendations based on opinions from “most similar” users, have been widely adopted in various applications. In spite of the overall success of CF systems, they encounter one crucial issue remaining to be solved, namely the cold-start problem. In this paper, we propose a method that combines human personality characteristics into the traditional rating-based similarity computation in the framework of user-based collaborative filtering systems with the motivation to make good recommendations for new users who have rated few items. This technique can be especially useful for recommenders that are embedded in social networks where personality data can be more easily obtained. We first analyze our method in terms of the influence of the parameters such as the number of neighbors and the weight of rating-based similarity. We further compare our method with pure traditional ratings-based similarity in several experimental conditions. Our results show that applying personality information into traditional user-based collaborative filtering systems can efficiently address the new user problem.