Application of collaborative filtering algorithm based on time decay function in music teaching recommendation model.

Area of Science:

• Educational Technology
• Computer Science
• Data Science

Background:

• Traditional teaching resource recommendation systems face challenges with data sparsity, scalability, and the cold start problem.
• Existing collaborative filtering (CF) methods often struggle to adapt to evolving user preferences over time.

Purpose of the Study:

• To enhance the accuracy and effectiveness of teaching resource recommendation systems.
• To address the limitations of traditional CF algorithms by incorporating temporal dynamics.

Main Methods:

• An enhanced collaborative filtering (CF) recommendation algorithm was developed, integrating a time decay (TD) function.
• The TD function, inspired by the human memory forgetting curve, was used as a weighting factor to calculate similarity and user preferences.
• This approach amplifies the weight of recent user interests, integrating short-term and long-term preferences.

Main Results:

• The proposed time decay collaborative filtering (TD-CF) algorithm achieved a Root Mean Square Error (RMSE) of 8.95 for 100 recommendations.
• This RMSE is significantly lower than that of the comparison model, indicating higher accuracy.
• The TD-CF model demonstrated superior performance across various recommendation scenarios, effectively utilizing music teaching resources and user characteristics.

Conclusions:

• The TD-CF algorithm effectively addresses data sparsity, scalability, and cold start issues in teaching resource recommendations.
• Incorporating a time decay function significantly improves recommendation precision by balancing short-term and long-term user interests.
• The enhanced algorithm offers a more accurate and personalized approach to recommending educational resources.