The integration of forensic accounting and big data technology frameworks for internal fraud mitigation in the banking industry

Abstract

The purpose of this study is to investigate the integration of forensic accounting and big data technology frameworks in relation to the mitigation of internal fraud risk in the banking industry. This study employed an explanatory research design involving the use of simulated data to mirror the situation in the banking industry. To this end, the big data analytical approach considered is machine learning that involves a neural network with two-layer feed forward, one hidden layer and five hidden neuron layers created to detect the presence of fraud and classify them into two, viz.: fraudulent and non-fraudulent activities. Both the input and output target samples are automatically divided into training, validation, and test datasets, while the confusion matrix is employed to visualise the percentages of correct and incorrect classifications. Furthermore, the clustering of the fraud indicators was also carried out to group them based on their similarities. The results obtained demonstrate the feasibility of neural networks in classifying internal fraud into three levels of risks and fraud detection. This is evidenced in the percentage of correct classification (95%) and misclassification (5%) obtained from the confusion matrix. The model also demonstrates the feasibility of clustering the potential red flags of internal fraud. This study provides an understanding into the attributes of internal fraud and a practical guided approach to implement an integrated forensic accounting and big data technology framework for internal fraud mitigation. The forensic accountant should ensure that the machine learning models are regularly updated with new datasets for automatic classification and clustering analysis. There is still scanty information regarding the integration of forensic accounting and big data technology for mitigation of internal fraud risk in the banking industry. Hence, it is envisaged that this study will contribute to the method, theory and practise of internal fraud mitigation.