Classification of Coronary Artery Lesions Based on XGBoost

XGBoost is an optimized distributed gradient enhancement library, which uses the gradient lifting method to build a strong learner after each iteration, and it is widely used in many classification and regression model scenarios. It is a scalable machine learning system of treeboosting, which can solve many data science problems quickly and accurately. Using retrospective analysis, 1552 patients were enrolled in the Department of Cardiology, Huashan Hospital affiliated to Fudan University from January 1, 2018 to December 31, 2020. A total of 1552 patients were treated and discharged from the hospital. According to the Gensini score, the patients were divided into four groups: low GSlow < 24, middle GSmid < 53, high GShigh < 120 and critical recombination GSdanger ≥ 120. In this paper, the data set of patients’ physiological indexes was analyzed by XGBoost algorithm, and the classification results of coronary artery lesions were predicted. The AUC values of the four groups were 0.88, 0.69, 0.87 and 0.96 respectively. The experimental results show that the algorithm has achieved good results in predicting the classification results of each group.