Loading...
Paper Type
ERF
Paper Number
1738
Description
Fraud has a significant impact on the financial industry. It not only results in financial losses but also reputation and customer trust, which cannot be quantified. Some of the recent reports show that just the financial cost of fraud is US$5 trillion. Due to its massive impact on organizations, it is vital for organizations to take the necessary actions to combat fraud. Fraud detection is a complex task and requires both technological, organizational and regulatory support. In this study, we concentrated on the technology part where we proposed a fraud detection action taking the fraud costs into account. Generally, machine learning algorithms do not take the associated costs into accounts. By using ensembles of artificial neural networks and logistic regression models, a sample-cost dependent fraud detection engine model is proposed. We believe the inclusion of cost and utilizing the ensemble learning method enables organizations to decrease their fraud detection costs significantly.
Recommended Citation
AKGUL, Mehmet, "A Cost-Based Fraud Detection System for Financial Sector" (2021). AMCIS 2021 Proceedings. 25.
https://aisel.aisnet.org/amcis2021/art_intel_sem_tech_intelligent_systems/art_intel_sem_tech_intelligent_systems/25
A Cost-Based Fraud Detection System for Financial Sector
Fraud has a significant impact on the financial industry. It not only results in financial losses but also reputation and customer trust, which cannot be quantified. Some of the recent reports show that just the financial cost of fraud is US$5 trillion. Due to its massive impact on organizations, it is vital for organizations to take the necessary actions to combat fraud. Fraud detection is a complex task and requires both technological, organizational and regulatory support. In this study, we concentrated on the technology part where we proposed a fraud detection action taking the fraud costs into account. Generally, machine learning algorithms do not take the associated costs into accounts. By using ensembles of artificial neural networks and logistic regression models, a sample-cost dependent fraud detection engine model is proposed. We believe the inclusion of cost and utilizing the ensemble learning method enables organizations to decrease their fraud detection costs significantly.
When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.