AI-Based Framework for Medical Cost Forecasting and Financial Transaction Monitoring in Healthcare

The growing digitization of healthcare systems necessitates accurate cost forecasting and robust protection of financial transactions. This paper proposes a novel AI-based framework that combines predictive analytics with cybersecurity techniques to address these dual challenges. Using the Medical Cost Personal Dataset comprising 1,338 records, we developed a Gradient Boosting Regressor that achieved a MAE of 1,873.92, RMSE of 3,912.14, and an R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> score of 0.88, outperforming other baseline models. In parallel, an Isolation Forest model trained on residual errors identified transaction anomalies with a recall of 0.89, demonstrating effective detection of irregular patterns. We also performed adversarial robustness testing, showing that manipulated features like smoker status can cause prediction shifts exceeding $4,000. These results underscore the potential of integrating machine learning and anomaly detection for both forecasting and safeguarding financial operations in healthcare. The proposed framework offers a scalable and efficient solution for building secure, data-driven financial infrastructures in medical systems.