A Recommendation Model for College English Digital Teaching Resources Using Collaborative Filtering and Few-Shot...

Area of Science:

• Educational Technology
• Information Science

Background:

• Effective management and recommendation of digital English instructional resources are crucial for modern education.
• Traditional collaborative filtering (CF) methods face challenges like data sparseness and scalability in recommendation systems.

Purpose of the Study:

• To design and implement an enhanced digital English instructional resource management recommendation system.
• To improve upon traditional CF algorithms to address scalability, data sparseness, and user cold-start issues.

Main Methods:

• The study employed a B/S structure with a hierarchical design architecture for system development.
• An improved collaborative filtering algorithm was developed and integrated into the system.
• Functional and non-functional requirements of a personalized educational resource recommendation system were analyzed.

Main Results:

• The enhanced algorithm achieved a recall rate of 96.37% and a recommendation accuracy of 95.31%.
• These results significantly outperform traditional CF methods, which showed a recall rate of only 6.37% in comparison.
• The system effectively addresses drawbacks of traditional algorithms, improving recommendation quality.

Conclusions:

• The developed system efficiently provides high-quality digital English teaching resources.
• The improved CF algorithm enhances recommendation accuracy and overcomes common challenges.
• This system facilitates easier access to valuable educational materials for both educators and students.