Friend Relationships Recommendation Algorithm in Online Education Platform

Jingda Kang, Juntao Zhang, Wei Song, Xiandi Yang

Abstract: With the development and application of online education, mining friend relationships between learners can improve interactive communication, enhance collaborative learning, and motivate learners to make mutual progress. However, existing methods only recommend well-known members through the number of likes and fans, which fail to consider the hidden interest points and content topics of members in the text. To address this problem, we propose an Evaluation Latent Dirichlet Allocation (EvaluationLDA) algorithm to recommend suitable friends for learners in online education. The EvaluationLDA algorithm clusters learners with similar learning interests to obtain Top-N friend recommendation sequences based on constructing learner document datasets, calculating learner similarity, and modeling the friend topic. We conduct experiments to demonstrate the effectiveness of our EvaluationLDA algorithm. The result shows that our EvaluationLDA algorithm can effectively recommend Top-N friend sequences in the online education platform.