A dynamic filtering recommendation algorithm based on topic

Collaborative filtering is one of the most successful and widely used recommendation technology in E-commerce recommendation systems. However, existing collaborative filtering algorithms face severe challenge of sparse user ratings and real-time recommendation. To solve the problems, a collaborative filtering recommendation algorithm based on topic is proposed. It divides the raw rating matrix into many sub-matrixes based on topic and forms clusters parallelly to reduce the data sparsity. Time-based data weight and acceleration-based data weight are proposed to dynamically reflect the change of user interests. The experimental results show that the novel algorithm can efficiently improve recommendation quality.