Phishing Detection Using Machine Learning Techniques

As there are many cyber-attacks that are going on in this world phishing is one of the most important cyberattack phishing attacks starts with fake messages which contain dangerous links or URLs that can be sent through emails and chat applications that messages target the victim and make him open the malicious link in this way hacker gets the important information about the target victim like passwords and login details. So, the user needs to be conscious about which is a malicious link, and which is a legitimate link. Machine learning can be one of the ways to classify the malicious links and the legitimate links by which we can stop the 95% of phishing attacks. This paper is about training machine learning models using phishing datasets to classify the URLs whether they are legitimate URLs or phishing URLs. Machine learning models used to train are KNN, Decision tree, Random Forest, Logistic Regression, CNN, RNN, and all the model's accuracy and all other evaluation metrics are compared and Logistic regression and CNN gave the highest accuracy of 95% and 96% respectively.