Academic performance evaluation based on knowledge sharing theory: feature importance analysis of dimensions in LightGBM model

Abstract

This study is based on the theory of knowledge sharing and uses the LightGBM model to evaluate students' academic performance, and deeply analyzes the importance of each dimension feature in the model. Specifically, the study successfully constructed the LightGBM model by collecting data from students' knowledge sharing scales and academic performance. In order to comprehensively evaluate the performance of the model, this study compared it with Random Forest and CatBoost models, and used feature importance analysis techniques to explore the impact mechanism of knowledge sharing willingness and ability on academic performance. The research results show that the LightGBM model performs better in interpretability. Among the various characteristics of knowledge sharing ability, question 6 and question 8 are particularly crucial, corresponding to students' knowledge retrieval ability and knowledge recognition ability, respectively, playing an important role in academic performance evaluation. This study has opened up new perspectives and methods for the field of academic performance evaluation, which helps educators to better understand the intrinsic relationship between students' knowledge sharing behavior and academic performance.