Cyber Vigilance, Management and Awareness: Unmasking Phishing Threats for the Business and Corporate Community

Abstract

The most popular threats in the universe of internet are phishing attacks. Emails, texts, and even fully functional fake websites are used by crooks to trick victims into disclosing their confidential data especially for business and cooperate community. The most coveted targets are large, financially sound, high-value companies like banks, government organizations, and e-commerce sites. The perpetrators imitate the websites and make them strikingly similar in order to perform a strike. The thoughtful problem on the internet is phishing URLs, and this study emphases on the possibility of machine learning that can be used to address it. The main objective is to use machine learning (ML) algorithms to examine and study the different characteristics of uniform resource locators (URLs) in order to identify phishing URLs. This study focuses on primary classification models which include K-nearest neighbors, Naive Bayes, “Random Forest, Decision Trees, Artificial neural networks (ANN)”, XGBoost and Support vector machines (SVM). This paper has two main goals: first, finding the best classifier among 4 options to detect phishing threads, and second, identifying the most effective feature selection method for phishing website datasets. Using one dataset for finding the high accuracy; the research has found Random Forest, XGBoost, SVM and Decision tree as the top classifier algorithms for detecting phishing websites.