A novel approach for collaborative filtering to alleviate the new item cold-start problem

Abstract

Recommender systems have been widely used as an important response to information overload problem by providing users with more personalized information services. The most popular core technique of such systems is collaborative filtering, which utilizes users' known preference to generate predictions of the unknown preferences. A key challenge for collaborative filtering recommender systems is generating high quality recommendations on the cold-start items, on which no user has expressed preferences yet. In this paper, we propose a hybrid algorithm by using both the ratings and content information to tackle item-side cold-start problem. We first cluster items based on the rating matrix and then utilize the clustering results and item content information to build a decision tree to associate the novel items with the existing ones. Considering the ratings on novel item constantly increasing, we show predictions of our approach can be combined with the traditional collaborative-filtering methods to yield superior performance with a coefficient. Experiments on real data set show the improvement of our approach in overcoming the item-side cold-start problem.