Application of a Dynamic Model for Student Recognition Based on Graph Attention Neural Network in Education

In student education, teaching students according to their aptitude is an important part of implementing the fundamental task of cultivating morality and talent. However, universities currently face problems such as single identification methods, late discovery, and single data objects when depicting different types of students. This article proposes a solution to these problems by constructing a student social network through course selection data, class information, dormitory information, one card consumption, and access control data. The student feature embeddings are constructed in databases such as school entry and exit information, and attention neural networks are trained using the results of heart-to-heart conversations as student labels. This approach effectively improves the recognition rate of different types of students, especially for special students, in a timely and accurate manner, to assist teachers in teaching students according to their aptitude.