PPGINF PÓS-GRADUAÇÃO EM ENGENHARIA DA INFORMAÇÃO FUNDAÇÃO UNIVERSIDADE FEDERAL DO ABC Phone: Not available http://propg.ufabc.edu.br/ppginfo

Banca de DEFESA: DANIEL HENRIQUE MIGUEL DE SOUZA

Uma banca de DEFESA de DOUTORADO foi cadastrada pelo programa.
STUDENT : DANIEL HENRIQUE MIGUEL DE SOUZA
DATE: 06/06/2023
TIME: 15:00
LOCAL: por participação remota (https://conferenciaweb.rnp.br/sala/claudio-30)
TITLE:

Fraud Detection in Credit Card Transactions: a Machine Learning Approach


PAGES: 187
BIG AREA: Engenharias
AREA: Engenharia Elétrica
SUBÁREA: Telecomunicações
SPECIALTY: Sistemas deTelecomunicações
SUMMARY:

In this paper, we analyze the problem of supervised machine learning and its applications in detecting fraudulent transactions in credit card payments. First, we discuss the concept of fraud in payment systems, its consequences, and the importance of detecting this type of transaction to mitigate risks. Next, we present supervised and unsupervised machine learning problems, as well as the main algorithms used (such as Bayesian Networks, Neural Networks, Decision Trees, and K-Means), their applications, computational implementation, and performance evaluation methods.

Next, we describe methodologies for combining machine learning algorithms such as classifier aggregation and the combination of supervised and unsupervised learning methods (Mixed Learning). As the main contributions of this work, we highlight the new CC-OR, CCK-VM, and CCK-OR algorithms, based on a new classifier aggregation function combined with the concept of Mixed Learning, as well as a variation of the K-Nearest Neighbors algorithm adapted for imbalanced data.

To evaluate the different estimators, we compared the main classifiers available in the literature operating individually and aggregated through majority voting, as well as the algorithms proposed in this work, evaluating their performance in detecting fraudulent operations. We conducted numerical simulations using programs written in Python, using both real and synthetic data, which showed gains in the use of the proposed methods when compared to the state-of-the-art of the field.


COMMITTEE MEMBERS:
Presidente - Interno ao Programa - 1671282 - CLAUDIO JOSE BORDIN JUNIOR
Membro Titular - Examinador(a) Interno ao Programa - 2334927 - ANDRE KAZUO TAKAHATA
Membro Titular - Examinador(a) Externo ao Programa - 1296761 - MARCELO BENDER PEROTONI
Membro Titular - Examinador(a) Externo à Instituição - DIOGO MARTINS GONCALVES DE MORAIS - FESA
Membro Titular - Examinador(a) Externo à Instituição - RODRIGO MAROTTI TOGNERI - FGV
Membro Suplente - Examinador(a) Externo à Instituição - RENATO MACHADO - ITA
Membro Suplente - Examinador(a) Externo à Instituição - DENIS GUSTAVO FANTINATO - UNICAMP
Notícia cadastrada em: 05/05/2023 13:50
SIGAA | UFABC - Núcleo de Tecnologia da Informação - ||||| | Copyright © 2006-2024 - UFRN - sigaa-1.ufabc.int.br.sigaa-1-prod