Performance Prediction Model Based on Multi-Layer Attention Mechanism for Programming Ability

In the course of Programming Design and Data Structure, how to improve students’ programming ability has become a focus of attention for teachers. Currently, OnlineJudge systems that can automatically judge students’ code submitted to a programming exercise problem has been widely adopted by a large number of universities for programming practice. If OnlineJudge systems can predict the score obtained by a student for a new problem, it can effectively help teacher know more about the actual programming ability of each student. Score prediction can also be taken as an important basis for improving teaching techniques. In this paper, we propose a performance prediction model based on multi-layer attention mechanism for students’ programming ability, which can predict the performance of student when solving a new programming exercise problem. Real-world dataset filtered from an OnlineJudge system is used to train and test the model. Experimental results show that the model proposed in this paper has better performance and interpretability comparing with other models. (Abstract)