Detecting Fraudulent Healthcare Insurance Claims using Ensemble Machine Learning and SMOTENC: A Review

Abstract

Healthcare insurance fraud has become a serious global issue and are creating financial losses and operational inefficiencies for insurance providers. As digital claim submissions are increasing, fraudulent activities are becoming more complex and harder to identify using traditional audit-based systems. This paper reviews existing research on detecting healthcare fraud using machine learning, emphasizing ensemble models and advanced class-imbalance handling methods like SMOTENC. It gives the effectiveness of algorithms such as Random Forest, Gradient Boosting, and XGBoost in recognizing evolving fraud patterns in healthcare datasets. The paper also emphasizes the role of data preprocessing, resampling, interpretability tools, and deployment frameworks to build reliable and scalable systems of fraud detection.