A complete literature review on financial fraud detection applying data mining techniques

Abstract

Financial fraud is defined as unlawful or criminal duplicity attempted to result to organisational or personal gain. It is a big threat to the economics of a firm, corporate sector, government or ordinary customers in the form of credit card fraud, insurance fraud, and financial statement fraud. Several approaches exist in the literature of financial fraud detection. But, due to the inefficiency of those approaches, researchers leverage data mining techniques to detect financial fraud. This paper aims to build a systematic academic review of financial fraud detection approaches based on data mining techniques in the recent years. In the practice, different data mining techniques namely: K-nearest neighbour, decision tree, fuzzy logic, logistic model, Bayesian belief network, naïve Bayes, Beneish M-Score model, Benfords law, Altman Z-score have been applied to improve the accuracy of fraud detection. In this paper, existing financial fraud detection techniques are compared with the advantages and limitations of the techniques.