Personalised higher education course recommendation system based on generative AI

The digital transformation of global higher education and personalised learning needs has driven the rise of recommendation systems. The traditional teaching mode is faced with the difficulties of overload of curriculum resources and low learning efficiency, and the recommendation system has become the key to solve the contradiction of ‘information overload-demand precision’. The purpose of this paper is to propose a personalised higher education course recommendation system based on generative AI to accurately match educational resources and learning needs, meet personalised learning needs, optimise resource allocation and promote educational equity. Then, the HD-TGNN-GAN model is constructed by fusing graph neural network (GNN), generative adversarial network (GAN) and reinforcement learning (RL) technologies. By utilizing graph learning to handle sparsity, HD-TGNN-GAN enhances personalized recommendations. For example, its accuracy is 0.92, F1-score is 0.85, recall is 0.90, and user satisfaction is 4.5/5. At the same time, its robustness test also performs well, and its new user click-through rate in the first month reaches 0.085. On the whole, the recommendation system proposed in this paper effectively improves the accuracy and personalisation of course recommendation, but it still needs to further optimise the algorithm and enhance data processing capabilities and privacy protection.