The majority of computing devices in the 21st century is deployed in the cloud, providing access to information and online services. However, phishing attacks continue to pose a risk to the confidentiality, integrity, and availability of information, exposing legitimate users to potential threats. Thus, despite previous efforts to combat phishing attacks and improve detection models, there is still room for further performance improvement. To address this issue, the study proposed a multi-level phishing attack detection scheme for the cloud environment. The experimental results via simulation show our proposed multi-level model for phishing detection can achieve 0.9997 each for accuracy and F1-measure, and 0.9998 for specificity and precision each, as well as 0.9996 and 0.0003 respectively. A comparison between the proposed approach and the benchmarked algorithm shows that the proposed approach performs significantly better. This indicates that the proposed model can be used by computing and application industries to prevent phishing attacks, and it demonstrates the effectiveness of integrating multiple models for research analysis and experimentation in the academic domain.
Key words: Precision, Cloud Computing, Accuracy, False Positive Rate, Recall
|
