A Method for Personalized Delivery of Teaching Resources in Vocational Colleges Based on Mining User Browsing Information

In the rapidly evolving domain of vocational education, the deluge of digital teaching resources necessitates an advanced personalized recommendation system to optimize learning experiences. Traditional recommendation algorithms struggle to grasp the dynamic nature of user interests and the intricate web of interactions between users and resources. Therefore, in this paper, an efficient approach combining dynamic user interest modeling with a graph-based recommendation system is proposed. Transformer architectures are utilized to capture temporal shifts in user preferences through feature engineering of browsing behaviors, click history, and preference settings. A graph neural network is employed to capture the complex relationships within a graph-based framework. It uses graph convolutional or attention networks to dynamically assign weights that reflect the influence of neighboring nodes. Experiments are conducted on real-world datasets to assess the performance of the proposed approach in terms of precision, recall, and $F$1-score compared to traditional systems. The proposed approach provides more accurate and personalized teaching resource recommendations in vocational education settings.