Early Detection of Hepatitis Disease Using Machine Learning Algorithms

Abstract

Hepatitis is an inflammation of the liver caused by viral infections, autoimmune disorders, or exposure to toxic substances. Hepatitis B and C are major public health concerns because they may progress to cirrhosis or liver cancer. In Indonesia, the transmission rate remains high, primarily through blood contact, unsterile needles, transfusions, and maternal delivery. Limited public awareness, coupled with the often asymptomatic nature of hepatitis, leads to delayed detection, which increases the risk of severe complications and mortality. Therefore, early detection is crucial to minimizing the disease burden.This study proposes a risk prediction model for hepatitis using non-laboratory clinical data and machine learning methods. Eight classification algorithms were compared, Naïve Bayes, K-Nearest Neighbor (K-NN), Random Forest, Support Vector Machine (SVM), Decision Tree, AdaBoost, XGBoost, CatBoost, and LightGBM. Model performance was evaluated through K-fold cross-validation using accuracy, precision, recall, F1-score, and AUC. The results show that the SVM with a linear kernel achieved the highest performance, with 87% accuracy and balanced F1-scores across all classes. The model successfully classified four categories, Acute Hepatitis, Chronic Hepatitis, Liver Abscess, and Parasitic/Viral Infections. These findings highlight the potential of machine learning to improve early detection of hepatitis effectively and efficiently.