A Novel Technique to Detect URL Phishing based on Feature Count

Abstract

The advent of internet access to people across the globe, increasing levels of connectivity, remote employment, de-pendence on technology, and automation have presented a rapid rise in cybercrime. There are numerous methods of cybercrime that are frequently used, with phishing being the most significant. Various algorithms have been proposed to classify malicious and legitimate URLs, blacklisting and whitelisting algorithms being the oldest, used for computer security. But these algorithms are time-consuming and involve human labor. Hence, machine learning algorithms were incorporated to improve the effectiveness of classifying URLs. Recent research includes detecting phishing URLs using machine learning classifiers but the accurate classification is a long-term goal. This paper focuses on evaluating models for classifying URLs using different feature counts and deriving a specific model to provide good enough classification with less time complexity.