The current study explores the use of various ensemble classifiers on the HCV dataset, incorporating an autoencoder for feature extraction to improve performance while reducing the number of features. A key objective of this research is to identify the most effective classifier that delivers optimal performance metrics. Accurate diagnosis of the disease stage in hepatitis C (HCV) patients is critical for timely and appropriate treatment, which can help prevent further deterioration of the patient’s health. To achieve this, machine learning techniques are applied to perform multi-class classification based on disease stage labels. The proposed approach employs the classification using different ensemble models upon features extracted using an autoencoder. The dataset consists of electronic health records of hepatitis C patients, provided by Kanazawa University, Japan. After preprocessing, ensemble classifiers such as Random Forest, AdaBoost, and XGBoost are employed for prediction followed by performance assessment. Given the multi-class and imbalanced nature of the dataset, performance is assessed using a comprehensive set of metrics, including accuracy, precision, recall, F1 score. Additionally, this research aims to enhance interpretability by identifying the most influential features affecting classifier performance. This is achieved using the SHAP (SHapley Additive exPlanations) tool for feature importance analysis.