Neural networks and cognitive diagnostic modeling for OBE-oriented curriculum association maps

In this paper, a graph neural network-based cognitive diagnostic model of course association is proposed to address the problem of modelling the association between course knowledge and competency goals in OBE.The model constructs a knowledge topology graph using course prior relationships, aggregates the information of neighbouring course nodes through graph neural networks to enhance the knowledge representation, and combines with the DINA model to achieve fine-grained diagnosis of learners' cognitive status.Experiments on the publicly available dataset MoocRadar show that the model in this paper achieves 86.7% in cognitive diagnosis ACC, which is 3.2% and 2.8% higher than the traditional IRT model and the plain DINA model, respectively, and the MSE is reduced to 0.103.The results show that the proposed model can effectively capture inter-course dependencies and support personalised teaching path recommendation under the OBE system.