The Emergence Threat of Phishing Attack and The Detection Techniques Using Machine Learning Models

With the increasing number of the internet users, cybercrime is increasing at a high rate. Among all the cyber- attack, phishing has now confirmed so successful and the number one attack vector. Throughout our investigation of phishing attack, we find that it is the most uses attack and used many ways to attack the targeted user. Attacks by phishing URL, phishing email and phishing websites are very popular way of phishing attack. But now-a-days with the increasing popularity of social media and online gaming, attackers are targeting this media for the phishing attack. Machine learning is giving a new era for both the attackers and the users who want to prevent this attack. In our work we build a twitter spear phishing bot using machine learning. We did experiment on the detection of phishing url, phishing email and phishing website. For the detection of phishing url we used various classifier and with the higher accuracy we focus on the timing to train the dataset. We found that XGBoost classifier gives the higher accuracy 94.44% and took less time. For the detection of phishing email, we used naïve bayes classifier and got the accuracy of 95.15%. In our website detection techniques, we used different classifier and found that Random Forest Classifier gives the higher accuracy of 96.80%