Nowadays, social networking is one of the most popular digital activities worldwide. This is the reason why a lot of cybercriminals try to exploit social media platforms, such as Facebook, Instagram, and Twitter, as source of attacks. This gave rise to the concept of Angler Phishing. During the previous years, a lot of researchers tried to deal with the problem of phishing on social media. However, nobody took care to deal with Angler Phishing, which is a slightly different concept. Therefore, I built a machine learning model capable of detecting support fraud accounts, which is faster than the algorithm currently used by Twitter itself.