Enhancing Hepatitis Patient Survival Detection: A Comparative Study of CNN and Traditional Machine Learning Algorithms

Abstract

Hepatitis patient survival prediction is a critical medical task impacting timely interventions and healthcare resource allocation. This study addresses this issue by exploring the application of a Convolutional Neural Network (CNN) and comparing it with traditional machine learning algorithms, including Support Vector Machine (SVM), Decision Tree, k-Nearest Neighbors (KNN), Gaussian Naive Bayes (GNB), and Gradient Boosting (GBoost). The research objectives include evaluating the algorithms' performance regarding confusion matrix metrics and classification reports, aiming to achieve accurate predictions for both "Live" and "Die" categories. The dataset of 155 instances with 20 features underwent preprocessing, including data cleansing, feature conversion, and normalization. The CNN model achieved perfect accuracy in hepatitis patient survival prediction, outperforming the baseline algorithms, which exhibited varying accuracy and sensitivity. These findings underscore the potential of advanced machine learning techniques, particularly CNNs, in improving diagnostic accuracy in hepatology.