Intelligent Hepatitis Typing: A Machine Learning Paradigm for Multi-Class Prediction

Abstract

In this study, we focus on the urgent need for a robust tool to diagnose and detect Hepatitis, while also predicting the life expectancy of patients suffering from the disease. We conducted a detailed comparative analysis using machine learning techniques and neural networks, assessing the performance of Support Vector Machines, K-Nearest Neighbours, and Artificial Neural Networks. Key metrics such as accuracy rate and mean square error were used to evaluate these models. Our goal is to provide valuable insights for selecting the most effective tool for Hepatitis diagnosis and prognosis, aiming to enhance patient care and outcomes. The research highlights the significance of early Hepatitis detection and classification by utilizing data-driven methodologies and comprehensive data pre-processing. Current machine learning systems often fail to accurately differentiate between types of Hepatitis, a gap our proposed system seeks to fill. By incorporating Logistic Regression, Random Forest, Decision Tree, and Support Vector Machine models, our system offers precise classification. It also features an intuitive interface for entering patient data, enabling the system to predict the specific Hepatitis type and suggest treatment options. By addressing the limitations of binary outcomes and small datasets, our work makes significant contributions to improving medical diagnosis and predictive capabilities, particularly in Hepatitis management.