Designing Educational Games to Foster Decision Experience in College Innovation and Entrepreneurship Programs

Educational games immerse learners in situational environments to boost engagement, which addresses gaps in monitoring knowledge mastery and sustaining enthusiasm in college Innovation and Entrepreneurship (IE) courses. This study constructs an internal logic mechanism linking IE education and behavior, leveraging Bayesian Networks (BN) for Probabilistic Reasoning (PR) to assess learners' knowledge mastery. Drawing on educational game frameworks for knowledge structure tracking, it proposes a college IE decision-experience game model enhanced by an improved Deep Knowledge Tracing (DKT) algorithm (integrating feature embedding and attention mechanisms). Results show the model's entrepreneurship education score and prediction accuracy both reach 95.6% (the highest among tested models), with all evaluation scale items demonstrating strong information coverage. Average indicator scores exceed 4, reflecting effective feedback. Students' adaptability to the game-based model is 2%-11% higher than to traditional teaching. The embedded real-time evaluation aligns learning performance with instructional goals, enabling strategy adjustment. IE education's value lies in fostering entrepreneurial awareness, capability, and willingness, while its function enhances practical skills like resource integration. The model improves learning enthusiasm, adaptability, and efficiency, offering insights for personalized college IE education design.