Identification of process-based fraud patterns in credit application

Fraud detection has become an important research topic recently. In a credit application, fraud can occur in forgery of documents or business processes. Fraud on the business process is known as Process-based Fraud (PBF). Previous studies proposed several detection methods of fraud in the business process model. This fraud detection includes analysis methods and an identification process. However, none of them proposed PBF identification, particularly identification of PBF attributes and pattern clearly, so its accuracy still needs further improvement. As identification of PBF attributes and PBF pattern is very important for the accuracy of PBF detection, this paper proposes an identification method for PBF detection. This PBF identification process consists of some attributes, those are skip sequence, skip decision, throughput time minimum, throughput time maximum, wrong resource, wrong duty decision, wrong duty sequence, wrong duty combine, wrong pattern and wrong decision. In this paper, PBF pattern is combined with a fuzzy set which consists of low, middle and high categories. This fuzzy set is implemented in order to improve the accuracy of PBF determination. PBF attribute and its pattern contribute to the process mining for detecting PBF.