Detecting Phishing Using a Multi-Layered Social Engineering Framework

Abstract

As businesses develop and expand with a significant volume of data, data protection and privacy become increasingly important. Research has shown a tremendous increase in phishing activities during and after COVID-19. This research aimed to improve the existing approaches to detecting phishing activities on the internet. We designed a multi-layered phish detection algorithm to detect and prevent phishing applications on the internet using URLs. In the algorithm, we considered technical dimensions of phishing attack prevention and mitigation on the internet. In our approach, we merge, Phishtank, Blacklist, Blocklist, and Whitelist to form our framework. A web application system and browser extension were developed to implement the algorithm. The multi-layer phish detector evaluated ten thousand URLs gathered randomly from the internet (five thousand phishing and five thousand legitimate URLs). The system was estimated to detect levels of accuracy, true-positive and false-positive values. The system level accuracy was recorded to be 98.16%. Approximately 49.6% of the websites were detected as illegitimate, whilst 49.8% were seen as legitimate.