Phishing Website Detection Using Machine Learning

Abstract

Phishing attacks continue to be a danger in our digital world, with users being manipulated via rogue websites that trick them into disclosing confidential details. This article focuses on the use of machine learning techniques in the process of identifying phishing websites. In this case, a study was undertaken on critical factors such as URL extension, age of domain, and presence of HTTPS whilst exploring the effectiveness of Random Forest, Gradient Boosting and, Support Vector Machines algorithms in allocating a status of phishing or non-phishing. In this study, a dataset containing real URLs and phishing URLs are employed to build the model using feature extraction. Following this, the various algorithms were put to the test on this dataset; out of all the models, Random Forest performed exceptionally well having achieved an accuracy of 97.6%, Gradient Boosting was also found to be extremely effective possessing strong accuracy and accuracy. In this study we also compared and discussed methods to detect a phishing site. Some features that affect detection performance include URL length, special characters and the focus on even more aspects that need further development. The new proposed method improves the detection accuracy of the phishing websites because machine learning techniques are applied, recall (true positive) increase, while false positive decrease. The results enrich the electronic security system, as they enable effective detection in real time mode. This study has demonstrated the importance of employing cutting-edge techniques to deal with phishing attacks and safeguard users against advanced cyber threats, thus laying the groundwork for innovation in phishing detection systems in the future