Liver disease prediction using W-LR-XGB Algorithm

Abstract

As a common disease in the population, liver disease is not easy to be found, and it is easy to develop into liver cirrhosis and liver cancer. This article proposes the W-LR-XGB algorithm to establish a prediction model for patients with liver disease to help doctors perform preliminary screening of patients with liver disease. This article uses the liver function records of 573 patients, uses the W-LR-XGB algorithm to establish a model to predict whether the patient has liver disease, and uses the accuracy rate, recall rate, and F1 score evaluation model. In the research, it is found that the W-LR-XGB algorithm is compared with the traditional machine learning algorithm. The model established by the CatBoost algorithm has better predictive ability and can make more accurate predictions for patients with liver disease. The accuracy rate is 0.83, the recall rate is 0.82, and the F1-score is 0.81.