An Interpretable LightGBM-Based Machine Learning Framework for Robust Multiclass Hepatitis C Virus Prediction

Hepatitis C Virus (HCV) is a blood-borne disease that causes chronic inflammation. This can then lead to liver cancer. It is usually spread through direct contact with blood. There is no vaccine for this disease. However, it can be treated with proper and effective treatment. Advances in technology have made machine learning (ML) an effective method of learning. There is no risk in this method, it helps in primary treatment. In our study, we used a dataset containing a total of 616 blood samples from patients to create a robust predictive framework. It was a multiclass dataset. In this study, we have used a total of$\mathbf{5}$ML models. Random Forest (RF), Extreme Gradient Boosting (XGB), Light Gradient Boosting Machine (LGBM), Stacking, Gradient Boosting (GB). We have used the$\mathbf{1 0}$-fold cross-validation method to improve the performance of the models. We found the highest accuracy among the models at 94.72 % (LGBM). To guarantee model transparency, we have added various visualization techniques including Correlation Heatmap, Violin Plot and Explainable AI (XAI) Frame Work like SHAP and LIME, Work Flow Diagram and many other types of Comparison Tables to better understand the entire scenario. The main objective of our study is to accurately predict (HCV) in advance and to accurately interpret that prediction.