Machine Learning-Based Security Mechanism for Detecting Phishing Attacks

Phishing attacks pose a serious threat to online users by impersonatina trustworthy entities and stealing sensitive data. This can result in reputational damage, financial loss, ransomware, or additional spyware outbreaks. Phishing scams take into consideration roughly 22% of all data violations, making it one of the most prominent cybercrimes, in accordance with the FBI's 2021 IC3 report. According to phishing email statistics, roughly 1.2% of all emails received are fraudulent, resulting in 3.4 billion phishing emails per day. One out of every 4,200 emails sent is almost certainly a phishing attempt. Cognitive methods, such as machine learning (ML) and deep learning (DL), are widely used in cybersecurity to extract information and detect possible intrusions and phishing attempts. This research investigates the procedure of employing machine learning methods for recognizing URLs that are used for phishing attacks. We have used the Pristine and Malicious URLs Dataset to train, test, and evaluate the performance of a set of algorithms to develop a robust binary class URL phishing attack detection mechanism. The obtained results show that the proposed mechanism outperforms the other existing competing mechanisms with 99.6% accuracy.