Text-Based Fraud Detection Using Deep Learning and Transformer-Based Hybrid Approaches

Wong, Chyi Keat (2026) Text-Based Fraud Detection Using Deep Learning and Transformer-Based Hybrid Approaches. Final Year Project (Bachelor), Tunku Abdul Rahman University of Management and Technology.

Text RDS_Wong Chyi Keat_Fulltext.pdf Restricted to Registered users only Download (1MB)

Abstract

Fraud detection has become a crucial concern in the 21st century, with fraudsters employing complex strategies to attack vulnerabilities in text-based communications. This study addresses the issue of text-based fraud by creating a hybrid deep learning and transformer-based strategy for efficiently detecting fraudulent actions. The proposed system, which uses advanced natural language processing (NLP) techniques and cutting-edge transformer models like BERT, would assess textual patterns, semantic meaning, and contextual relevance to detect probable fraud. Furthermore, explainable AI (XAI) methods will be used to improve robustness and transparency. The study's goal is to create a scalable system capable of identifying fraud in real time, with an emphasis on increasing financial security and digital protection for organizations and individuals. This innovation promises to advance cybersecurity practices and provide valuable contributions to safer online communication.