University Enrollment Plan Configuration Optimization Model: A Big Data-Driven Approach

The research on the optimization method of enrollment plan configuration based on data-driven and artificial intelligence has always been a hot topic in the field of higher education teaching reform. Due to the unclear linkage patterns between enrollment, education, and employment, various universities have not yet established well-quantified models, making it difficult to form a stable enrollment plan configuration optimization mechanism. This article uses real enrollment data of all majors from 2020 to 2023 at a specific university deployed by the Ministry of Industry and Information Technology. It analyzes the linkage effects at each foused stage, prioritizes 7 indicators using RF importance and Birnbaum importance, and conducts Pearson correlation analysis and multifactor variance analysis on past data. Then, using the enrollment planning number as the target variable, it establishes Ridge Regression, SVR, GBDT, RF, and XGBoost regression models, while performing five-fold cross-validation to evaluate model performance in terms of R-Square. Experimental results show that the GBDT regression model has a R-Square as high as 92.2%, and this model can provide reliable predictions for the enrollment plans of various majors in various provinces at the university in 2024.