Aiming at the pain points of traditional financial case teaching, such as lagging case update, low matching degree and insufficient resource utilization, this paper proposes and constructs a financial case intelligent push and personalized teaching recommendation system based on deep learning. The dynamic knowledge representation framework is introduced systematically and innovatively, and external variables such as industry prosperity index and policy heat are integrated into case modeling to improve the timeliness of cases; Construct a portrait of a multimodal learner, fuse eye tracking data with the LSTM (Long Short-Term Memory) time series model, and realize the fine evaluation of the learner' s ability; Design a closed-loop mechanism of "recommendation-feedback-optimization" to support personalized teaching in the whole process before, during and after class. The experiment is based on the real data of Advanced Financial Management course in a university. The results show that the system is significantly better than the traditional static recommendation model in HR@5 and NDCG@5, and the knowledge mastery is improved by 56%, and the user satisfaction is overall leading. The research verifies the feasibility and effectiveness of deep learning technology in intelligent finance teaching, and provides theoretical support and practical path for building an adaptive and extensible personalized teaching system.
Yan Lin (Sun,) studied this question.