Prediction Model of Hypertension Complications Based on GBDT and LightGBM

Abstract

Abstract Complications caused by hypertension include heart failure, stroke, arteriosclerosis, etc. The prediction of hypertension complications is a hot issue, and it is difficult to predict it from a medical perspective. In this study, we aim to establish a prediction model of hypertension complications based on machine learning and data mining. We first proposed a GBDT-based feature selection method, which can screen out medical indicators that affect the hypertension complications. On this basis, we established a hypertension complications prediction model based on LightGBM. The results show that after 10-fold cross-validation and comparison analysis, the accuracy, F1 and AUC of the prediction model are 0.9189, 0.8888, and 0.9233 respectively, which are significantly better than other machine learning models. Therefore, the proposed method can accurately predict hypertension complications, so as to provide effective clinical auxiliary diagnosis for doctors and help them take preventive measures to reduce the impact of hypertension complications.