Academic Performance Prediction Model Based on Educational Similarity

Academic performance is a direct assessment indicator of student learning achievement that aims to predict future learning outcomes based on historical student behavior data. Existing academic performance methods tend to ignore the influence of students' home education on academic performance. Therefore, to address the above issues, this paper proposes an academic performance prediction method based on educational similarity(APPES). We construct the similarity of students with the same home education background and use a graph convolutional neural network (GCN) to integrate it with student characteristics for student academic performance prediction. Our method outperforms existing methods through experiments on real student datasets.