A new method for occupational fraud detection in process aware information systems

Today, companies have turned to use fraud detection methods to reduce their financial losses that have been arisen in this way. Thus, Process aware information systems are vulnerable to insider frauds. Flexibility in these systems gives the opportunity for fraudsters to commit illegal activities. Strict security controls on these systems at runtime reduces their flexibility. Moreover, the frequent changes in these systems make inefficient the ordinary fraud detection methods and it remains as a challenge for organizations. In this paper, we propose a new fraud detection method that uses both statistical information about system's log and process model mined from it to detect fraudulent instances. Our method reduces false positive rate and supports loop, parallel and selection structures in processes. The experimental results show effectiveness of the approach as it represents value of more than 0.8 for F-measure.