Providing a Model for Detecting Tax Fraud Based on the Personality Types of Corporate Financial Managers using the Neural Network Approach

One of the management measures to reduce tax liabilities is non-payment of taxes through tax fraud. Because personality factors may play a role in explaining tax ethics, examining personality traits and aspects of tax fraud can help to better understand the factors that influence tax decisions. The main purpose of this study is to provide a model for detecting tax fraud based on the personality types of corporate financial managers using the neural network approach. The statistical population of the study consists of all financial managers of listed and non-listed companies in 2020 who are not exempt from taxes. The information about financial managers was collected through a questionnaire and analyzed using SPSS software version 21 and MATLAB software version 2015. The results indicate that the neural network model designed with 10 neurons in the hidden layer with an accuracy of 79.5% has the ability to detect tax fraud committed by financial managers of companies. Also according to the results of the regression model test, personality traits of neuroticism, extraversion, flexibility and agreement have a positive and significant effect on tax fraud and the personality trait of conscientious has a negative effect on tax fraud.