Web-based phishing attacks detection with machine learning methods

In this thesis, firstly the threats on information security are described; methods of defense against these threats, and recommendations are presented. Two applications in order to identify the phishing website were performed. In the first application, classification application has been performed for the detection of phishing website by Artificial Neural Network model. 5-fold cross-validation test has been applied to the created networks with the first application. The average classification accuracy was achieved as 90.61%, and the highest classification accuracy was achieved as 92.45%. In the second application, classification application has been performed for the detection of phishing website by Extreme Learning Machine model. In the second application, 10-fold cross-validation test has been applied. The average classification accuracy was achieved as 95.05%, and the highest classification accuracy was achieved as 95.93%.