Day surgery has been developed in China for over 30 years. However, the admission decisions for day surgery patients are still mainly based on expert experience, brainstorming, and institutional recommendation, and lack scientific admission criteria and universality. West China Hospital of Sichuan University has implemented methods such as semi-supervised learning based on heterogeneous data, to construct a more objective and accurate patient admission model based on large amounts of diagnosis and treatment data. This initiative aims to assist the country and hospital in reducing medical costs and alleviating the acute problem of the current contradiction between supply and demand of medical resources. It also seeks to optimize the utilization and allocation of bed resources, reduce the cost of patient management, enhance the theoretical research on patient admission methods in day surgery in China, and provide reference and inspiration for other hospitals in the day surgery industry in China.