Year: 2025
Type: article
Abstract: : Intoday’s digitalenvironment frauddetectionisamajorproblemthat impactsfinancialservices, e-commerce, banking and insurance.Because online transactions are growing so quickly, scammers are always coming up with new ways to get around established security measures. Because of their restricted feature extraction capabilities and dependence on predefined rules, traditional rule-based and machine learning approaches frequently fall short in identifying complexfraudpatterns. Inthisstudy,weinvestigatetheuseofdeep learningtechniquesforfraud detection, suchasLongShort-TermMemory(LSTM) networks, ConvolutionalNueralNetworks (CNN), and Recurrent Neural Networks (RNN).To learn transactional patterns and spots irregularities instantly, these algorithms are trained onactualfinancialrecords. The performance of deep learning models and conventional fraud detections techniques is compared in the study using important assessment criteria like accuracy, precision, recall and F1-score. Our results show that by detecting intricate correlations in transactional data traditional methods dramatically increase fraud detection rates. This studyalso addresses issues like data imbalance, processing costs, and model interpretabilitythat arise when using deep learning for fraud detection. We provide several ways to get over these obstacles and improve the scalabilityand effectiveness of fraud detection systems. The study’s findings demonstrate how deep learning may be used to improve fraud preventions systems and guarantee safer online financial transactions.
Author: Suraj Gupta
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W4409477005