Machine Learning Based Architecture for Cyber Threat Classification and Early Warning Detection

Abstract

As Internet communication has been increased to exchange of data and this sensitive information will be accesed by various unsuspecting users. Our main goal is to control phishers in attacking critical data such as usernames, passwords, and bank account details. In networks we face various threats. Domain phishing involves in attacking email recipients and to reveal their account details through links in user emails text for masquerading and getting legitimate registrars. URL which is designed to promote scams, attacks, and fraud. When a user clicks those urls that leads ransomware downloads and our information is attacked performing a cybercrimes. Hackers are mostly exploit new ttechniques like social engineering, phishing, Cybersecurity professionals are now seeking reliable and robust methods for detecting phishing websites. Our proposed project uses ML to detect various phishing URLs and attacks through analyzing various features of legitimate and phishing URLs. In our proposed we have used Support Vector Machine, Decision Tree, Gradient Boosting, Logistic Regression. The main ggoal this project is to identify phishing URLs and determine the most effective ML algorithm by comparing accuracy rates, false positives, and false negatives of each algorithm.