A Mobile-Based Personalized Physical Education Recommendation and Learning Path Optimization Model

Abstract

With the advancement of mobile technologies, personalized physical education (PE) has emerged as a critical component of intelligent education. However, existing recommendation models commonly suffer from limited adaptability, insufficient multimodal data integration, and poor alignment between learning paths and learners’ real-time states. To address these challenges, this study proposes a three-layer architecture for personalized recommendation and learning path optimization that integrates multi-source perception, dynamic cognition, and real-time optimization. The core innovations of the proposed model include: (1) a lightweight multi-source heterogeneous data fusion module designed for efficient on-device processing; (2) a dynamic tri-state assessment model incorporating skill mastery, fatigue level, and interest level to achieve accurate real-time perception of learners’ states; and (3) a duallayer optimization mechanism based on online learning to enable end–cloud collaborative learning path optimization. This study provides a novel technical paradigm for mobile technology–enabled intelligent physical education, offering significant theoretical contributions and practical application value.