A Comparative Analysis on the Relative Success of Mixed-Models for Financial Statement Fraud Risk Estimation

Abstract

Loses which are caused by financial statement fraud (FSF) revealed the necessity of early warning system in fraud detection. In this context, many models have been improved. The level of success of these models on accurate estimation of financial statement fraud is proved by some empirical studies. Success level of the models has been discussed in the literature. Main purpose of this study is to reveal relative success of the models which are used in order to estimate FSF by considering the findings in the literature. The findings of this study show that variables of estimation of FSF include variations and also there is not any consensus on this issue in the literature. Additionally, it is concluded that artificial neural network models are more successful than other models in estimation of FSF.