Research on XGboost academic forecasting and analysis modelling

Abstract

The analysis and evaluation of student achievement is an important part of teaching work and an important part of school routine management. Scientifically analysing and evaluating students’ academic achievements can not only enable teachers to accurately grasp the students’ learning status, but also enable students to understand their own learning situation, and also provide necessary analysis for teaching management and improving teaching. In order to evaluate students’ learning situation comprehensively, objectively and reasonably, this paper uses XGBoost algorithm to classify and evaluate students’ performance based on statistical analysis of basic data, and establishes a performance evaluation model. For the curriculum relevance, the student’ s performance data is statistically compiled according to the statistical knowledge. The subjective and objective structural entropy weight method is used to classify the characteristic importance results, and finally the relevant courses of the completed courses are obtained. For the results of the unfinished course using the completed course results, the XGBoost method is used to predict each student’ s grades.