Know URL: Machine Learning-based Chrome Extension for Detection of URL Phishing

Phishing is a con game that scammers/imposters use to collect personal information from unsuspecting users. It’s a major problem that is spreading year after year very rapidly, especially increased by 220 phishing incidence during the pandemic period. Traditionally, phishing detection has relied on either a blacklist of known phishing URLs or a heuristic evaluation of the features of a suspected phishing website. The heuristic algorithm uses trial and error to determine the threshold for classifying harmful and benign links. The disadvantages of this method include its lack of accuracy and flexibility to new phishing URLs. To overcome the drawbacks of traditional phishing detection methods, the proposed system provides an efficient machine-learning solution for detecting phishing and fraudulent URLs. In the proposed system, model is trained using several light weight algorithms such as logistic regression, KNearest Neighbors, Support Vector Machine, and Naïve Bayes and select the highly accurate and efficient model i.e., Support Vector Machine (SVM) with an accuracy of 95.7% with less load time compared to other existing models. Using the efficient model, a chrome extension phishing detection mechanism has been developed to alert the users about the illegitimate web sites and preventing the phishing attack, thereby scales down the cyber crime and cyber crisi that arises year by year.