Phishing attacks are one of the most damaging attacks for internet users. Detecting these attacks is one of the main challenges in the internet security due to their lack of unpredictable nature. Machine learning techniques are suitable methods to detect these attacks. The accuracy of these methods is highly dependent on the features of the data. This paper proposes a hybrid feature selection approach based on Bee algorithm and Logistic Regression method, which detects phishing attacks accurately by selecting an appropriate feature subset. The results show that the proposed approach has been able to increase the detection accuracy of phishing attacks by 1.23%, using a random forest algorithm, in comparison to previous works with 6.66% less features.
