Intelligent Detection Approaches for Securing E-Banking Platforms against Phishing Websites.

Abstract

E-banking has introduced fresh innovations in finances because of the new conveniences, efficiencies, and global access. Conversely, the fast spread of e-banking has increased phishing attacks, which are a cyber-attack on banking sites. This paper concentrates on the intelligent detection of phishing websites to protect e-banking systems, particularly the assessment and comparison of algorithms on intelligent detection. The UCI Phishing Dataset URL, content, and behavioral features were diverse and applied in training and testing different machine-learning models. Support Vector Machines, Random Forest, Neural Networks, XGBoost, and Hybrid performance models were evaluated in terms of accuracy, precision, recall, F1-score, and AUC. Although SVM showed a mediocre detection risk, the Random Forest and Neural Networks proved to be significantly more reliable, whereas XGBoost exhibited the highest performance due to its accuracy and scalability of performance. All tests yielded the most reliable results, and hybrid systems achieved the highest metrics and performance. This is revealed the most reliable detection and control systems of phishing threats in e-banking system since phishing frauds hugely make use of vulnerabilities in e-banking. In online banking systems, phishing frauds are best controlled by hybrid systems and intelligent detection systems. Some of the practical refinements provided by this study are improving fraud prevention, customer confidence, and regulatory compliance within financial organizations. The investigations of the future must focus on creating mechanisms of fraud detecting in real-time, using larger and more diverse data sets, and more adaptable and adaptable learning systems that can evolve with phishing attacks to sustain the digital banking protection mechanisms.