A collaborative filtering algorithm based on rating distribution
Deng Shuangyi, Liang He, Xia weiwei · 2008
Collaborative filter algorithms are one of the most successful recommender technologies in the world, and have been widely adopted in E-commerce. However, these approaches always suffer from poor prediction quality problem. We analyzed the rating distribution of dataset, and dig out that most of the users are interested on several specific topics, which show the user’s true interest. So we propose a new collaborative filtering algorithm based on rating distribution (BRDCF). Finally, we experimentally evaluate our approach and compare it with classical collaborative filtering methods; the results demonstrate the effectiveness of our approach.