An Effective Technique for Early Cirrhosis Prediction Using Machine Learning

The chronic liver disease cirrhosis is becoming a major worldwide health concern because of its rising rates of morbidity and mortality. Using 17 clinical criteria, the “Cirrhosis Patient Survival Prediction” study aims to predict the survival outcomes of patients with liver cirrhosis. The aim of the ML models developed for this investigation is to accurately classify patients into three primary groups: Censored (C) patients whose outcomes were not fully observed; Death (D) patients who died as a result of complications during the study period; and Censored due to Liver Transplantation (CL) patients whose successful outcome was believed to increase survival. With the use of this study, our models improve our understanding of the prognosis of patients with cirrhosis, allowing for more customized treatments and overall improved healthcare. This study looks into the prediction of mortality in cirrhosis patients using machine learning models. The accuracies achieved by the models are 72.6(%) by SVM, 75(%) by Logistic Regression, 95(%) by Neural Network, 70(%) by Decision Tree and 80.95(%) by Random Forest. It offers a thorough rundown of methods in three primary domains: validation and assessment, model training and development, and feature extraction and selection. By merging several machine learning techniques, this survey seeks to offer significant insights into the evolving field of predictive analytics for cirrhosis outcomes. It is meant to be a useful tool for researchers, physicians, and other healthcare professionals who are figuring out how to combine machine learning and liver disease treatment.