HEALTHCARE FRAUD DETECTION USING MACHINE LEARNING ENSEMBLE METHODS

Abstract

Healthcare fraud can lead to significant financial losses and disrupt patient care. Detecting fraud in medical claims is a challenging task due to the volume of data and changing fraud patterns. Machine learning (ML) techniques, especially boosting techniques, have shown great success in improving the accuracy of fraud detection. Boosting algorithms such as Adaptive Boosting (AdaBoost), Gradient Boosting, and Extreme Gradient Boosting (XGBoost) improve predictive performance by combining weak classifiers into a robust model. This paper has develop a framework by using ensemble learning based ML models like XGBoost, LightGBM, and found that they performs well as compare to other methods. The method also uses SMOTE for resolving class imbalance problem in dataset. The work has been performed on Medicare claim dataset provided by Kaggle. This paper investigates the use of a special technology for medical fraud detection and compares its results with other models. Experimental results show that the developed system can improve the accuracy of classification, recall, and correctness, becoming a powerful tool for medical care fraud detection. Future research can focus on the integration of deep learning and descriptive AI techniques to improve fraud detection and further explanation.