MOOC Resources Recommendation Based on Heterogeneous Information Network

Aiming at the problem that the existing MOOC recommendation mechanism cannot meet the dynamic and diversified learning needs of different individuals, a MOOC resource recommendation model based on heterogeneous information network is proposed. First by capturing MOOC platform of the heterogeneity between multiple entities in building its corresponding heterogeneous information network, and then through the node level attention and meta-path level fusion of attention, will learn to the user and the knowledge incorporated into the extended matrix factorization framework, to predict user preferences for knowledge, to carry on the personalized recommendation service. Experimental results show that this model has better recommendation performance than other commonly used models, and effectively solves the problem of personalized recommendation for learners.

Automated summary

A MOOC resource recommendation model based on heterogeneous information network is proposed to address the problem of the existing MOOC recommendation mechanism not meeting the dynamic and diversified learning needs of different individuals. By capturing the heterogeneity between multiple entities in the MOOC platform and utilizing node level attention and meta-path level fusion of attention, this model incorporates user and knowledge into an extended matrix factorization framework to predict user preferences for knowledge and provide personalized recommendation service. Experimental results demonstrate that this model outperforms other commonly used models and effectively solves the problem of personalized recommendation for learners.