Explainable AI (XAI) on Health Insurance Claim Prediction

This research endeavors to explore the XGBoost Machine Learning algorithms through the lens of Explainable Artificial Intelligence (XAI) and to furnish elucidations for the derived outcomes. In this manuscript, we have proposed a methodology for predicting health insurance claims. The proposed framework entails addressing the classification challenge utilizing the XGBoost algorithm, succeeded by an exposition of the anticipated results through XAI methodologies. The anticipated medical insurance claims are further elucidated through model-agnostic techniques, specifically LIME and SHAP. User experience improvement and establishment of computational model-user trust represent the primary reasons why this investigation is necessary. The research tools enable domain experts to confirm predictive model accuracy while they inspect key influencing features so they can use their specialized knowledge base.