Exploring Financial Fraud Detection: A Comprehensive Analysis and Implementation of Machine Learning with Artificial Neural Networks

Author Bertha C. Ori
Co-Author Cletus I. Ori, Lilian O. Ezekiel
DOI https://ww

Country : Nigeria
Subject :

Keywords : Financial fraud; fraud detection; machine learning; data mining; support vector machine (SVM), artificial neural network (ANN).

Abstract

Financial fraud, characterized as deceptive strategies aimed at securing financial gains, has emerged as a widespread threat to companies and organizations worldwide. Traditional methods like manual verifications and inspections are not only imprecise but also incur high costs and time consumption in identifying fraudulent activities. The rise of artificial intelligence has paved the way for intelligent machine learning approaches to efficiently detect fraudulent transactions through the analysis of extensive financial data. This paper seeks to offer a systematic literature review (SLR) that methodically examines and consolidates existing literature on machine learning (ML)-based fraud detection. To conduct this review, the artificial neural network approach was employed to demonstrate fraud detection procedure with 70 % of sample data for training, 15% for testing and 15% for validation. Numerous studies were collected through specified search strategies from popular electronic database libraries. Following the application of inclusion/exclusion criteria, a considerable number of articles were thoroughly examined, synthesized, and analyzed. The review provides an overview of prevalent ML techniques employed in fraud detection, the most common type of fraud addressed, and the evaluation metrics utilized. The scrutinized articles revealed that support vector machine (SVM) and artificial neural network (ANN) are popular ML algorithms employed for fraud detection, with credit card fraud being the most frequently addressed fraud type using ML techniques. The paper concludes by highlighting key issues, identifying gaps, and delineating limitations in the field of financial fraud detection. Additionally, it suggests potential areas for future research in this domain.

Download