PhishNull: Enhancing Cyber Hygiene Through Supervised Machine Learning

The research paper addresses the pervasive issue of phishing within the realm of Internet security, acknowledging its persistence despite the advancements in antivirus and technical safeguards. Focusing on combating this online scam, the study delves into two primary methodologies: Black Listing and Machine Learning. Opting for a Machine Learning and heuristic-based approach, the thesis conducts a comparative analysis of various Machine Learning algorithms, including Logistic Regression, alongside ensemble algorithms such as Adaboost and Gradient Boost. While initial expectations leaned towards ensemble algorithms yielding superior results, the outcomes revealed a nuanced reality. Although ensemble algorithms demonstrated promising predictive capabilities, their performance did not surpass expectations.