Supervised learning methods for fraud detection in healthcare insurance

Abstract

Fraud in the healthcare system is a major problem whose rampant growth has deeply affected the US government. In addition to financial losses incurred due to this fraud, patients who genuinely need medical care suffer because of unavailability of services which in turn incur due lack of funds. Healthcare fraud is committed in different ways at different levels, making the fraud detection process more challenging. The data used for detecting healthcare fraud, primarily provided by insurance companies, is massive, making it impossible to audit manually for fraudulent behavior. Data-mining and Machine-Learning techniques holds the promise to provide sophisticated tools for the analysis of fraudulent patterns in these vast health insurance databases. Among the data mining methodologies, supervised classification has emerged as a key step in understanding the activity of fraudulent and non-fraudulent transactions as they can be trained and adjusted to detect complex and growing fraud schemes. This chapter provides a comprehensive survey of those data-mining fraud detection models based on supervised machine-learning techniques for fraud detection in healthcare.