Abstract

In today’s digital world, fraud detection has become an important part of financial security. This study explores and compares the performance of different machine learning models in identifying fraudulent transactions using the IEEE-CIS Fraud Detection dataset. Techniques such as Random Forest, Gradient Boosting, Deep Neural Networks, and Logistic Regression were evaluated. The dataset was pre-processed using SMOTE to balance the classes and improve model sensitivity to fraud cases. Each performance of the model was assessed using accuracy, precision, recall, and F1-score. The Random Forest model achieved the highest overall performance with an F1-score of 99.23

Library of Congress Subject Headings

Fraud investigation; Forensic accounting; Big data; Data mining

Publication Date

7-17-2025

Document Type

Thesis

Student Type

Graduate

Degree Name

Professional Studies (MS)

Department, Program, or Center

Graduate Programs & Research

Advisor

Ioannis Karamitsos

Campus

RIT Dubai

Plan Codes

PROFST-MS

Share

COinS