Artificial Intelligence: To Fraud Detection and Analysis for Insurance Claims in USA using machine learning algorithms

This research examines how AI could transform US healthcare insurance claim processing. It tries to find the best machine-learning(ML) model for health insurance claim prediction, saving insurance companies money. Various metrics assessed six ML algorithms' health insurance claim prediction performance. SVM, DT, RF, LR, XGBoost, and KNN are explored. Key measures are used to evaluate performance in research. A feature importance analysis reveals the key variables that affect insurance claim prediction. With 79% and 77% R-squared values and the lowest prediction errors, the XGBoost and RF models outperformed the other techniques. The feature importance analysis shows that smoking, BMI, and blood pressure are crucial to insurance claim prediction. These findings underscore the need to include these variables in insurance policy and pricing formulations. This study shows that AI, particularly the XGBoost model, can improve healthcare insurance claim processing precision and efficiency. Identifying critical variables and reducing prediction errors indicate the possibility for significant cost reductions and AI integration into healthcare insurance operations. This study supports the use of AI for healthcare insurance process optimization and data-driven decision-making.