Article Published In Vol.10 (July-Aug 2022)

AI-Driven Fraud Detection and Prevention Framework for Cloud-Based Banking Systems

Author : Archana Chaluvadi, Winner Pulakhandam, Visrutatma Rao Vallu and R. Hemnath

Fraud detection in cloud-based banking systems has become increasingly critical as financial transactions continue to move online, leading to sophisticated fraudulent activities. Existing methods face challenges such as limited adaptability to new fraud patterns, imbalanced datasets, and scalability issues, which hinder fraud detection and efficient processing of large transaction volumes. This paper presents an AI-driven fraud detection system designed to overcome these challenges by leveraging deep learning techniques and cloud-based infrastructure. The system begins with the collection of transaction data, user behavior logs, and account details, followed by data preprocessing using Min-Max scaling for numerical data and one-hot encoding for categorical variables. Next, feature extraction using Continuous Wavelet Transform (CWT) is applied to the pre-processed data, capturing temporal and frequency patterns. These features are used to train a Variational Autoencoder (VAE) model to detect anomalies. The trained model is deployed to a cloud-based platform, ensuring scalability and fraud detection. Experimental results show that the system achieves an accuracy of 99.51%, precision of 98.92%, sensitivity of 98.77%, specificity of 99.12%, F-measure of 98.96%, and throughput of 800 transactions per second. This work provides the enhancement of fraud detection systems by integrating advanced AI techniques and cloud infrastructure, offering a robust, scalable, and efficient solution for large-scale banking environments.

Keywords: Fraud Detection, Banking Systems, Cloud Computing, Deep Learning, Variational Autoencoder and Continuous Wavelet Transform.