The deep learning-based physical education course recommendation system under the internet of things

Area of Science:

• Computer Science
• Educational Technology
• Data Science

Background:

• Traditional physical education course recommendations lack personalization and accuracy.
• Existing systems struggle with data sparsity and the cold start problem.

Purpose of the Study:

• To propose a deep learning (DL)-based physical education course recommendation system.
• To enhance recommendation accuracy and personalization by integrating Internet of Things (IoT) technology and DL.
• To address data sparsity and cold start issues using Generative Adversarial Network (GAN) models.

Main Methods:

• Utilized IoT devices (smart bracelets, smart clothing) to monitor real-time physiological and environmental data.
• Captured students' social interactions for socially oriented course recommendations.
• Integrated IoT data with academic data for optimized course matching.
• Employed the Regularization Penalty Conditional Feature Generative Adversarial Network (RP-CFGAN) model to handle data sparsity and cold start problems.

Main Results:

• The proposed DL-based system demonstrated strong performance in TopN evaluations.
• Significant enhancements were observed compared to traditional recommendation models.
• The integration of IoT and GAN models improved the understanding of student needs for personalized recommendations.

Conclusions:

• The combination of IoT technology and GAN models offers a robust solution for personalized physical education course recommendations.
• The system effectively addresses key challenges in recommendation systems, improving accuracy and user satisfaction.
• Future work includes exploring advanced regularization, ensuring user privacy, and expanding system applicability.