Liver Cirrhosis Prediction Using Machine Learning Algorithms

Abstract

Liver cirrhosis is a highly infectious blood-borne illness that is often asymptomatic in its early stages. Consequently, identifying and curing patients at an early stage of the ailment is difficult. As the disease spreads to its advanced stages, diagnosis and treatment become even more difficult. The goal of this research is to propose an artificial intelligence system utilizing machine learning algorithms conservatively created to support physicians in diagnosing liver cirrhosis at an early stage. To estimate the likelihood of a liver cirrhosis infection, it was decided to construct a model for training produced machine learning algorithms that use a variety of physiological variables. Three models for accurate prognostication have been developed with different training, linking three different sets of physiological variables and machine learning algorithms built on LR, SVM, DTC, GB and RFC. The algorithm that performed the best in this challenge was Random Forest, which had an accuracy of almost 86.74%. The approach was developed using the publiclyaccessible Liver Cirrhosis data source. The models used in this study had a significantly higher accuracy than those used in previous studies, indicating their increased dependability. Their resilience has been demonstrated by several model comparisons, and the research study may be used to select the scheme.