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

Banca de QUALIFICAÇÃO: DANIEL HENRIQUE MIGUEL DE SOUZA

Uma banca de QUALIFICAÇÃO de DOUTORADO foi cadastrada pelo programa.
DISCENTE : DANIEL HENRIQUE MIGUEL DE SOUZA
DATA : 30/03/2021
HORA: 14:00
LOCAL: por participação remota
TÍTULO:

Fraud Detection in Credit Card Transactions: a Machine Learning Approach


PÁGINAS: 96
GRANDE ÁREA: Engenharias
ÁREA: Engenharia Elétrica
SUBÁREA: Telecomunicações
ESPECIALIDADE: Sistemas deTelecomunicações
RESUMO:

In this work we propose the analysis of the problem of supervised learning and some of its applications to the financial market, focusing on the recognition of fraudulent transactions in payments via credit card.
To this aim, first, we describe some concepts about payment method fraud, its consequences, and the importance of recognizing this type of transaction for risk mitigation. Then, we describe supervised learning problems, in which the objective is the estimation of categorical (classification) and numerical (regression) variables from labeled data, and the unsupervised learning problem, in which the intention is to build clusters on unlabeled data given dissimilarity measures. We first perform a literature review addressing the concepts of each learning method, the main mathematical models for solving each of the problems (Bayesian Networks, Neural Networks, Decision Trees, K-Means, among others), main applications, computational implementation and performance evaluation methods.
Next, in the practical part, aiming to evaluate the behavior of the different estimators, we make a comparative analysis between the main classification models, analyzing their performance in recognizing fraudulent transactions for payments via credit card. All numerical simulations are performed using functions written in the Python programming language, and the performance of the tested models is measured via tests performed on real data and computationally generated data. At the end of the work, we describe the results of the performed simulations, make some final considerations and a proposal for continuity this study.


MEMBROS DA BANCA:
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 - 3008017 - DENIS GUSTAVO FANTINATO
Notícia cadastrada em: 01/03/2021 09:28
SIGAA | UFABC - Núcleo de Tecnologia da Informação - ||||| | Copyright © 2006-2021 - UFRN - sigaa-1.sigaa-1