Student score prediction method based on hypergraph neural network

Abstract

The invention provides a student score prediction method based on a hypergraph neural network, which is used for solving the problem in the prior art that multi-source heterogeneous behavior data of students cannot be analyzed, so that academic scores cannot be accurately predicted. The method comprises the following steps: firstly, extracting multi-source behavior characteristics according to multi-source heterogeneous data of students, then, carrying out sensitivity analysis on the multi-source behavior characteristics of all the students to obtain influence characteristics of each behavior, and then, constructing a multi-source behavior hypergraph by utilizing the influence characteristics; and finally, inputting the multi-source hypergraph H of the student and a multivariate influence feature X formed by splicing the influence features of the four behaviors into a deep network, and predicting student scores. According to the Ms-HGNN method provided by the invention, the influence of a behavior mode constructed by multi-source behavior characteristics on student scores is considered from the perspective of groups, the student behavior multivariate association is flexibly represented, and certain interpretability is given while the model prediction accuracy is improved.

Description

technical field [0001] The invention mainly relates to the fields of educational data mining, hypergraph neural network and deep learning, in particular to a student performance prediction technology based on a multi-source hypergraph neural network. Background technique [0002] Using behavioral data to model student behavior and realize the analysis and evaluation of learning effectiveness has been a lot of research results in the academic circle. Due to the limitations of technology and tools, traditional theoretical research on learning behavior can only discuss those behaviors that can be directly measured and observed. However, for the research on the implicit correlation of learning behavior, there has been a lack of first-hand event data, and a systematic theoretical analysis system has not yet been formed. In recent years, with the construction of the campus information system, a large amount of data related to student learning has gradually accumulated, which provi...

AI Technical Summary

Problems solved by technology

[0007] In order to solve the problem that the existing technology cannot analyze the multi-source heterogeneous behavior data of students, and thus cannot accurately predict the academic performance, the present invention provides a performance prediction method based on the multi-source hypergraph neural network analysis of student behavior patterns. The hypergraph is introduced into the field of learning behavior modeling, and the hypergraph is constructed by extracting student behavior characteristics to capture the hidden high-order correlations in the multi-source heterogeneous behavior data of students, and the neural network is used to train node embedding features to achieve the goal of predicting grades. Purpose

Images