A Hybrid Adaptive Learning Model for Mathematics Education: Integration of WFCAL and Q-Learning

With the rapid development of artificial intelligence and big data technologies, personalized learning systems have become a key focus in education, aiming to tailor educational content to individual students' needs. However, existing systems still face challenges in fully adapting to students' diverse learning styles and behaviors, often failing to provide optimal learning paths and recommendations based on comprehensive student data. To address these shortcomings, this study proposes a personalized mathematics education system based on adaptive learning algorithms, integrating Weighted Feature Clustering Adaptive Learning (WFCAL), Weighted Knowledge Point Similarity Recommendation, and Q-learning Path Optimization. This novel approach dynamically evaluates students' knowledge mastery and adjusts the learning content and path accordingly, ensuring that each student follows the most efficient and personalized learning trajectory. The system’s key advantages lie in its ability to leverage both cognitive and behavioral data to optimize learning outcomes and enhance student engagement. Experimental results demonstrate that the proposed model significantly outperforms traditional methods in terms of knowledge mastery prediction accuracy, task completion rate, learning progress rate, and knowledge retention. Specifically, our model achieves higher learning efficiency and effectiveness across two publicly available datasets, ASSISTments and MATH, confirming the superiority of the adaptive learning approach. In summary, this research contributes to the field of personalized education by providing a robust and adaptive learning system that not only enhances learning outcomes but also offers valuable insights for future advancements in intelligent educational systems.