Economics at your fingertips  

Deep Neural Networks for Detection of Credit Card Fraud

Angelin Lalev ()
Additional contact information
Angelin Lalev: D.A. Tsenov Academy of Economics, Svishtov, Bulgaria

Conferences of the department Informatics, 2019, issue 1, 263-274

Abstract: The purpose of this paper is to present preliminary results from ongoing study, concerning application of Deep Neural Networks (DNN) to the detection of credit card fraud. The main approach is testing the change of different archi-tectural parameters of the neural network. The influence of the change on the performance and convergence of the network is measured. Some of the chosen parameters are the depth of the network, the width of each layer and the chosen activation functions in each layer. The main results of this study indicate that deep neural networks with moderate number of hidden layers in combination with techniques of oversampling are performing markedly better in detecting fraudulent transactions than networks with only one hidden layer. The used sample dataset is imbalanced, which means that the DNN training on the dataset has tendency to overfit. The results of this study are useful for other researchers and practitioners who try to analyze real datasets in order to detect card frauds.

Keywords: credit card fraud; deep neural networks; DNN; imbalanced dataset (search for similar items in EconPapers)
JEL-codes: C8 (search for similar items in EconPapers)
Date: 2019
References: Add references at CitEc
Citations: Track citations by RSS feed

Downloads: (external link) ... .years%20263-274.pdf (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link:

Access Statistics for this article

Conferences of the department Informatics is currently edited by Vladimir Sulov

More articles in Conferences of the department Informatics from Publishing house Science and Economics Varna Contact information at EDIRC.
Bibliographic data for series maintained by Vladimir Sulov ().

Page updated 2020-03-29
Handle: RePEc:vrn:katinf:y:2019:i:1:p:263-274