DEVELOPMENT OF AN INTELLIGENT AND SECURED MODEL FOR A CLOUD-BASED DETECTION OF ADVERSARIAL ATTACKS USING ARTIFICIAL NEURAL NETWORK

To identify and mitigate adversarial attacks, this paper explores the development of an intelligent and secure model for a 5G cloud-based facility utilising Artificial Neural Networks (ANN) and sophisticated statistical approaches. The research uses the Structured System Analysis and Create Methodology (SSADM) to analyse and create the suggested solution. Principal Component Analysis (PCA) and Mahalanobis distance were used in conjunction with a multi-layered neural network model to detect unusual packet behaviour and facilitate effective threat identification. The study makes use of an actual 5G cloud dataset with many traffic types, and preprocesses the data using techniques like feature extraction and imputation.Achieving an exceptional ROC score of 0.98673 and a detection accuracy of 99.6%, the detection model was assessed utilising critical performance indicators including Positive Predictive Value (PPV), False Discovery Rate (FDR), accuracy, and Receiver Operating Characteristic (ROC) analysis. The suggested ANN-based model outperformed the state-of-the-art algorithms in a comparative analysis, showing a 98.67% detection rate. The results demonstrate how well the approach works to improve cybersecurity for 5G cloud networks by offering a strong defence against hostile attacks.