Graph data mining and evaluation system optimization of educational quality based on graph convolutional network

Abstract

In educational globalization, the evaluation and improvement of educational quality have become the core issue of educational reform in various countries. With the acceleration of educational informatization, how to use massive educational data to realize the accurate evaluation of educational quality has become a research hotspot. Based on this background, this study puts forward the optimization research of educational quality graph data mining and evaluation systems based on graph convolutional networks. Through in-depth analysis of educational quality data in a middle school, this study constructs an educational quality graph data model and applies graph convolutional network technology to train and optimize the model. The experimental data covers multiple dimensions such as student achievement, teacher information, and school facilities, totaling thousands of sample points. It is found that the method based on a graph convolutional network can effectively identify the key factors affecting the quality of education, which is helpful in improving the prediction accuracy and interpretation ability of the evaluation system. The experimental results show that compared with the traditional evaluation methods, the method proposed in this study has significantly improved the evaluation accuracy, reduced the error rate by about 15%, and can more comprehensively reflect the actual situation of education quality. This research not only provides new technical means for educational quality evaluation but also provides scientific and effective data support for educational decision-makers, which has far-reaching significance for promoting educational equity and high-quality development.