Fraud Detection Using Neural Networks: A Case Study of Income Tax

Abstract

Detecting tax fraud is a top objective for practically all tax agencies in order to maximize revenues and maintain a high level of compliance. Data mining, machine learning, and other approaches such as traditional random auditing have been used in many studies to deal with tax fraud. The goal of this study is to use Artificial Neural Networks to identify factors of tax fraud in income tax data. The results show that Artificial Neural Networks perform well in identifying tax fraud with an accuracy of 92%, a precision of 85%, a recall score of 99%, and an AUC-ROC of 95%. All businesses, either cross-border or domestic, the period of the business, small businesses, and corporate businesses, are among the factors identified by the model to be more relevant to income tax fraud detection. This study is consistent with the previous closely related work in terms of features related to tax fraud where it covered all tax types together using different machine learning models. To the best of our knowledge, this study is the first to use Artificial Neural Networks to detect income tax fraud in Rwanda by comparing different parameters such as layers, batch size, and epochs and choosing the optimal ones that give better accuracy than others. For this study, a simple model with no hidden layers, softsign activation function performs better. The evidence from this study will help auditors in understanding the factors that contribute to income tax fraud which will reduce the audit time and cost, as well as recover money foregone in income tax fraud.

Digital taxation, artificial intelligence and Tax Administration 3.0: improving tax compliance behavior – a systematic literature review using textometry (2016–2023)

“…The repeated mention of “taxpayers” reinforces the centrality of taxpayer engagement and behavior in the discourse. The inclusion of terms such as “neural networks” and “fraud detection” indicates an exploration of AI’s role in combating tax fraud (Murorunkwere et al , 2022), while “systematic review” and “innovation” underscore methodological approaches and the drive for novel solutions within the field (Bassey et al , 2022).…”

“…Central to this network is “tax compliance,” which connects directly with “tax administration” and “digitalization,” highlighting the transformative role of digital advancements in tax systems and compliance approaches. Surrounding “tax compliance,” we see “artificial intelligence” and “blockchain technology” as significant, suggesting their potential to reshape tax administration and combat tax fraud (Murorunkwere et al , 2022; Raikov, 2021).…”

“…Key findings, drawn from research like that of Bellon et al (2022) and Murorunkwere et al (2022), indicate substantial enhancements in tax compliance and operational efficacy through the implementation of digital invoicing and AI-powered fraud detection systems. Such advancements are particularly beneficial for SMEs, which are more susceptible to compliance issues.…”

The Moderating Effect of Auditor Size in the Relationship between Using Artificial Intelligence Techniques and Fraud Risk Assessment

“…Artificial Neural Networks (ANN) are defined as a collection of algorithms that use approaches similar to those found in the human brain (Murorunkwere et al, 2022). ANN algorithms are used by (Işık et al, 2023) to test the accuracy of detecting fraudulent transactions and achieved 99.7981% accuracy in detecting fraud.…”

Simultaneous Classification and Regression for Zakat Under-Reporting Detection

“…The authors claim that the interpretability of the detected outliers is performed through the training of explainableby-design surrogate models over outliers validated internally. Recently, in [30], the authors coupled Artificial Neural Networks with a real dataset to detect factors related to income tax fraud. Their approach was designed to reduce time, effort, and cost taken by auditors in the manual identification of cases to be audited.…”