Enhanced Phishing Website Detection: Leveraging Random Forest and XGBoost Algorithms with Hybrid Features

Phishing technique is used by hackers or
attackers to scam the people on internet into giving
private details such as login credentials of various
profiles, social security numbers (SSNs), banking
information, etc. Attackers disguise a webpage as an
official legit website. Blacklist or whitelist, heuristic, and
visual similarity-based anti-phishing solutions are unable
to detect zero-hour phishing assaults or newly created
websites. Older methods are more complex and not
suitable for day-to-day scenarios since they rely on
external sources such as search engines. As a result,

finding newly constructed phishing websites in a real-
time context is a significant hurdle in the field of

cybersecurity. This paper presents a hybrid feature-based
anti-phishing approach that nullifies these problems by
extracting characteristics from URL and hyperlink data
that is only available on the client side. Also, a brand-new
dataset is created for experiments employing well-liked
machine-learning classification techniques.