Medical Insurance Medication Anomaly Detection based on Isolated Forest Proximity Matrix

Abstract

With the continuous development of the medical insurance system and the continuous increase in the number of insured persons, medical insurance fraud has also intensified. Traditional fraud detection methods are no longer suitable for the needs of medical insurance fraud review. Aiming at the problem of drug reselling in medical insurance fraud, this paper uses the proximity matrix of the isolated forest to improve the LDOF algorithm, and proposes a new medical insurance medication anomaly detection algorithm. Through comparative experiments, our algorithm not only improves the accuracy of LDOF, but also improves the computational efficiency, and shows good performance in the application of actual medical insurance data sets.