Robust Classification of Hepatitis C: Leveraging Random Forest and Hybrid SMOTE-ENN for Imbalanced Dataset

Hepatitis is a major global health issue, mostly because it affects important organs like the liver. The Hepatitis C virus (HCV) is regarded as a substantial worldwide health concern owing to multiple factors. Among the global population of 185 million individuals afflicted with HCV, approximately 350,000 people die from diseases caused by HCV. This study focuses on utilizing classification techniques to detect hepatitis using a dataset sourced from the UCI (University of California Irvine ) Machine Learning Repository. The dataset encompasses various clinical and demographic features essential for hepatitis diagnosis. However, an initial analysis identified a data imbalance, with certain classes having considerably fewer samples than others, potentially impacting the classification model’s performance. To address this imbalance, the Synthetic Minority Over-sampling Technique (SMOTE) combined with Edited Nearest Neighbors (ENN) was employed to increase the representation of minority class samples and improve data quality. The classification was conducted using the Random Forest Classifier algorithm, which is well-regarded for its ability to manage large and complex datasets. The results indicated that the model achieved 100% accuracy on the training data and 97.5% on the testing data. These outcomes highlight the effectiveness of the SMOTE-ENN approach in managing data imbalance and the reliability of the Random Forest algorithm in classifying hepatitis. This study contributes to advancing improved diagnostic techniques for hepatitis and can be adapted for future clinical use.