Application Model of Machine Learning in Intelligent Recommendation Systems for Educational Resources

View at publisher Summarise with VIRAT
Authors: Li Qianling, Zheng Zhiyuan, Zheng Lian, Su Chunyan
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

This paper proposes a multimodal resource recommendation algorithm (DCG-AMR) based on a dynamic cognitive graph and attention mechanism. The algorithm employs a four-layer architecture consisting of "data layer - feature layer - recommendation layer - application layer": The dynamic cognitive graph updates the user's cognitive state in real time, while a customized attention mechanism is used to assign feature weights, simultaneously achieving deep fusion of multimodal resource features. Experiments were conducted on the EdNet dataset (containing 102,000 users, 53,000 multimodal resources, and 10.86 million interaction records), comparing four algorithms: CF, CBR, NCF, and KGRec. Results show that DCG-AMR achieves Precision@10 of 82.3% (a 25.3% improvement over CF) and NDCG@15 of 81.5% (a 12.2% improvement over KGRec), with a 19.6% improvement in learning effectiveness and a 77.8% resource completion rate. Ablation experiments demonstrate the effectiveness of the core module (precision drops by 7.7% after removing DCG). Stability tests show that precision fluctuates by less than 3% when the number of users varies. Scalability tests show that precision only drops by 1.2% when live streaming resources account for 100%. This algorithm effectively improves recommendation accuracy and user experience, providing support for optimizing domain-specific intelligent recommendation systems.

Fields of Study & Subjects

Recommender Systems and Techniques Online Learning and Analytics Intelligent Tutoring Systems and Adaptive Learning

Publication details

Journal / source
IEEE International Conference on Power, Electronics and Green Energy (ICPEGE)
Pages
879–884
Publication date
28 January 2026
Language
EN