Privacy Preserving Secure and Efficient Detection of Phishing Websites Using Machine Learning Approach

Abstract

One of the major worldwide crimes, phishing entail the burglary of the user's secretive information. Phishing websites frequently target the websites of business, institutions, government, and cloud storage space providers. While using the internet, the best parts of individuals are not aware of phishing assaults. Several phishing techniques now in use don't inefficiently address the troubles caused by email attacks. To combat software attacks, hardware-based phishing techniques are now deployed. The proposed effort concentrated on a three-stage spoofing series attempt for precisely identifying the difficulties in a material manner because of the increase in these types of problems. Uniform resource locators, circulation, and internet content based on phishing attack and non-phishing website strategy aspects were the three input variables. A dataset from previous phishing campaigns is gathered to apply the suggested phishing attack technique. Realistic phishing cases were found to provide a higher level of accuracy in phishing detection mechanisms and zero- day phishing attack. The categorization accuracy for phishing recognition using three dissimilar classifiers was indomitable to be 95.18 percent, 85.45 percent, and 78.89 % for NN, SVM, and RF, correspondingly. The findings indicate that a method based on machine learning works the best for phishing detection.