Prediction of New Student Registration Rate in Private Universities Based on the Random Forest Model

The New Student Registration Rate (NSRR) of private universities in China has a direct impact on the operation and development of the institutions. In order to enhance the prediction of NSRR in private universities, this study first identifies the key features of pre-admitted student information (AS) using the Random Forest algorithm. Then, it combines the AS information with the classic Random Forest algorithm to propose a Random Forest prediction model (P_Ror) for NSRR. The Random Forest algorithm is employed to train the AS dataset from 2022. The prediction model (P_Ror) achieved accuracy, precision, recall, and F1 scores of 94.23%, 94.16%, 94.23%, and 94.89%, respectively, on the test set. The prediction accuracy of the model was significantly higher than that of classic algorithms such as KNN, CART, and Naive Bayesian. The experimental results demonstrate that the proposed model effectively predicts NSRR in private universities and can offer valuable insights for resource planning.