Detection and Classification of Malicious Websites

Abstract

Phishing is quite possibly the most appealing technique used by attackers in the point of taking the individual subtleties of unsuspected individuals. Phishing sites are essentially tricks that are used by data fraud hoodlums and fakes. They use spam, fake sites made to look like the first sites, email, and direct messages to trick somebody into sharing significant information, like passwords and secret information. New enemies of phishing techniques are coming out each day, yet attackers think of new ways by focusing on all the new enemies of phishing techniques. So there is an earnest requirement for new strategies for the expectation of phishing sites. The paper portrays the correlation models in the classification of phishing sites for expectation utilizing distinctive Machine learning models. Different models are used for predicting which model gives the best exactness in phishing site classification. All the information is classified as either Benign for substantial Websites or Phish as Phishing Websites. Results have generated that show RF gives the best performance on this dataset for the classification of phishing sites.

Malicious Websites Classification Using Machine Learning Techniques: A Survey Paper

“…To detect dangerous websites from the OpenPhish domain, the top-performing classifier is chosen. Shubhankar, Siddhartha Bhaumik, and Prakash Biswagar [11], The approach for training the model is described, in which the seven machine learning algorithms are combined to produce the most adequate model for classifying benign and malicious websites. The most popular models are taken in for prediction at that stage when the models are split and then training is completed for each of the models for the various phases.…”