International Journal of Computer Applications
Foundation of Computer Science (FCS), NY, USA
|
Volume 187 - Issue 13 |
Published: June 2025 |
Authors: Krati Lodha, Katib Showkat Zargar |
![]() |
Krati Lodha, Katib Showkat Zargar . AI-Powered Detection of Financial Deception: Uncovering Credit Card Fraud. International Journal of Computer Applications. 187, 13 (June 2025), 39-46. DOI=10.5120/ijca2025925257
@article{ 10.5120/ijca2025925257, author = { Krati Lodha,Katib Showkat Zargar }, title = { AI-Powered Detection of Financial Deception: Uncovering Credit Card Fraud }, journal = { International Journal of Computer Applications }, year = { 2025 }, volume = { 187 }, number = { 13 }, pages = { 39-46 }, doi = { 10.5120/ijca2025925257 }, publisher = { Foundation of Computer Science (FCS), NY, USA } }
%0 Journal Article %D 2025 %A Krati Lodha %A Katib Showkat Zargar %T AI-Powered Detection of Financial Deception: Uncovering Credit Card Fraud%T %J International Journal of Computer Applications %V 187 %N 13 %P 39-46 %R 10.5120/ijca2025925257 %I Foundation of Computer Science (FCS), NY, USA
The surge in digital financial services has created new vulnerabilities to fraud, requiring advanced detection systems. Conventional fraud identification methods struggle with real-time processing, particularly when analyzing severely imbalanced datasets. This study introduces a multi-faceted AI framework combining tree-based boosting algorithms (LightGBM, XGBoost, CatBoost) with neural computation to improve fraud identification. Utilizing the creditcard.csv dataset containing 284,807 transactions where only 0.17% represents fraudulent activities, 1, 16 specialized techniques were implemented rebalancing approaches and parameter optimization to enhance detection performance. Results demonstrate that tree-based boosting approaches excel in precision metrics, lowering false alerts, while neural computation achieves superior sensitivity and discrimination capability 3, 4, 5. Specifically, XGBoost reached 88.17% precision with 97.25% area under curve, 4 CatBoost maintained balanced performance indicators, 5 and the neural architecture delivered 82.65% sensitivity with 97.95% discrimination capability 49. These outcomes illustrate how computational intelligence enhances financial security protocols, reducing unauthorized activities and minimizing institutional risk exposure.