A Machine Learning Model for Detecting Fraud in Health Insurance

Document Type : Original Article

Authors
Abstract
Fraud has become one of the major challenges in the insurance industry, where big data and machine learning are actively employed to address this issue. This study investigates health insurance fraud using machine learning algorithms and data mining techniques. The proposed innovative approach enhances fraud detection accuracy and contributes to reducing healthcare costs. We adopted a supervised learning method applied to automobile claim data from an anonymous insurance company. Our objective is to present an approach that improves the interpretability of artificial intelligence outcomes. The findings demonstrate that the Random Forest algorithm delivers the best performance based on the K-measure criterion and achieves the lowest RMSE score. Moreover, Random Forest provides the most balanced trade-off between recall and precision, with a detection rate of 23.8% of fraudulent cases among the total identified frauds.
Keywords