AI-Driven Fraud Detection and Transaction Security Framework for Digital Banking Systems
Year: 2026
Type: conference-paper
Abstract: This paper introduces the AI-based fraud detection and transaction security framework of digital banking systems, which applies the XGBoost, which is a powerful supervised machine learning algorithm. The issue of real-time identification of fraudulent actions and the need to maintain a smooth user experience have become the major problem in digital banking systems. The suggested framework applies the XGBoost that is trained on the transactional data including the transaction amounts, the types of merchants, and user behaviors. The model can correctly identify legitimate and fraudulent transactions by determining the risk score of every transaction. XGBoost is effective in detecting outliers and anomalies, which are indicative of fraud by using such properties as transaction velocity and behavioral patterns. This technique is accurate, precise and recallable and is therefore applicable in real-time application in detecting fraud on a large scale. The framework also incorporates the infrastructure of the banking system in place so that cases of fraud detected are raised and examined without having to interfere with the normal operations of the user to a considerable degree.
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7203679862