A Student Engineering Competence Level Prediction Method Based on RNN-GRU and Dynamic Bayesian Network

To achieve effective feedback and proficiency tracking of students in contemporary educational settings, we propose a methodology that integrates the RNN-GRU model and dynamic Bayesian networks for predicting students’ engineering competence level based on their online learning behaviors. The approach employed utilizes the RNN-GRU model to analyze students’ progression in algorithm development skills and algorithm implementation ability, resulting in the generation of a historical record of their learning features. These features are then employed to train a dynamic Bayesian network model, facilitating the prediction of students’ engineering competence level. After the amount of data reaches a certain amount, it can compare the previous models such as RNN and LSTM to obtain higher accuracy.