Research on Prediction of Liver Disease Based on Machine Learning Models

Abstract

Liver disease is a disease that has attracted much attention in the world. Liver diseases such as cirrhosis and liver cancer are the common causes of death in the world. Many liver diseases have no obvious symptoms in the early stage of onset, so they are easily overlooked by people. Treatment for liver illnesses is crucial and relies heavily on early diagnosis and management. This study assessed the effectiveness of various machine learning approaches for the identification of liver disease due to the high cost and complexity of the diagnostic process. This research used five machine learning models to predict the presence of liver disease based on a patient's medical records using an Indian liver patient record dataset. The dataset was prepared for model training through data preprocessing and analysis, including handling missing values, encoding categorical variables, and normalizing features. Five machine learning algorithms were evaluated, with Random Forest emerging as the highest performing model with an accuracy of 73.56% on the test set. This study contributes to the field by demonstrating the potential of machine learning to accurately predict liver disease, aiding in early diagnosis and treatment.