Optimization of Innovation and Entrepreneurship Education Models and Course Recommendation in Vocational Colleges Using LSTM Neural Networks

In this paper, we propose a personalized course recommendation framework for vocational education, leveraging a novel Attention-LSTM neural architecture. The model integrates a temporal attention mechanism and a dynamic time-window strategy to effectively capture students’ behavioral sequences and evolving learning interests. To enhance personalization, an interest similarity fusion module is introduced to combine historical preferences with recent engagement patterns. The framework is trained and evaluated on large-scale interaction data collected from an online learning platform, and compared against several baseline models including UserCF, ItemCF, MLP, and vanilla LSTM. Experimental results demonstrate significant improvements in Precision, Recall, F1-score, and NDCG metrics. Ablation and robustness studies further confirm the critical role of each proposed component in boosting performance and adaptability. The proposed system offers a scalable and intelligent solution for personalized content delivery in vocational education environments.