Fraudster Detection Based on Label Propagation Algorithm

Abstract

In recent years, healthcare fraud has happened frequently. Doctors, patients and pharmacies are involved, and these various subjects cooperate with each other to form a complicated fraud relationship. Healthcare fraud has caused great damage to the safety of the health insurance fund. However, fraudster detection problem has not been solved properly at present. Therefore, Fraudster Detection Based on Label Propagation Algorithm (FDBLPA) approach is developed to find patient fraudsters in a good accuracy. The construction of the patient-drug heterogeneous network makes the solution to healthcare fraud more diversified; New Speaker-listener Label Propagation Algorithm(NSLPA) improves the problem of random initialization of Speaker-listener Label Propagation Algorithm(SLPA), and has higher accuracy and efficiency. The main steps of our approach are following: (1) constructing a patient-drug heterogeneous network; (2) calculating the weight of the network; (3) using NSLPA algorithm for label propagation, and partitioning patients and drugs into corresponding communities; (4) finding patient fraudsters through community comparison. Experimental results suggest that the method we developed has a good performance in detecting patient fraudsters.