Analysis and optimization of student learning paths based on CRNN and sequential data

Abstract

The analysis and optimization of student learning paths have become increasingly critical in modern education, as they enable personalized learning experiences and improved academic outcomes. However, existing approaches often struggle to effectively model the temporal dynamics and knowledge point relationships inherent in learning behaviors. To address this challenge, this study proposes a novel framework that integrates Convolutional Recurrent Neural Networks (CRNN) for sequential feature extraction, Transformer models for knowledge point association modeling, and Reinforcement Learning (RL) for dynamic path optimization. Experimental results demonstrate significant improvements in learning completion rates (15% increase) and test scores (12% improvement) compared to baseline methods. The findings highlight that the integration of CRNN, Transformer, and RL provides a robust and scalable solution for personalized learning path analysis, offering actionable insights and adaptive recommendations to enhance student learning experiences and outcomes. This framework not only advances the field of learning analytics but also paves the way for more effective and inclusive educational technologies.