A learning resource recommendation method combining user sequential interaction with collaborative filtering

Wenjuan Niu, Zhendong Niu, Shiping Tang, Zhi Huang, Wei Wang, Xi Li, and Yaxin Chen


Elearning, collaborative filtering, sequential interaction, learning resource recommendation, learning material


Recommender system can make personalized predictions of resources for users with their learning history automatically. Collaborative filtering is one of the most widely used algorithms in this field. Although various works of collaborative filtering has been researched in e-learning, few of them notice the influence of sequential interactions among users. In this paper, we propose a novel collaborative filtering method by using the sequential interaction information of users. The proposed method consists of four steps: (1) fetching sequential interactive information from comments and replies; (2) computing the interaction influence degree among users with data from step (1); (3) filling the sparse user-item matrix with influence data; and (4) applying the new filled matrix to user-based collaborative filtering to find similar users to recommend. The experiment results on TED dataset show that the proposed method outperforms user-based CF and item-based CF on both precision and recall.

Important Links:

Go Back