Optimal Detection of Phising Attack using SCA based K-NN
Rajalakshmi Shenbaga Moorthy, P. Pabitha
Abstract
Phising is a dangerous social engineering cyber crime that aims to steal personal information of the user by sending spoofed emails. There are several applications of machine learning used to detect phising attacks. However, when a security model is built, the attacker tries to breach through it. Thus the question of optimal prediction of phising attack is still unanswerable. In this work, we integrated sine cosine algorithm, a metaheuristic population based technique with K-Nearest Neighbor (SCAK-NN) for predicting phising attacks optimally. The proposed SCAK-NN is compared with Decision Trees and Naïve Bayes. The results are satisfactory when comparing with various performance metrics such as accuracy, F-measure, True positive Rate (TPR), False Positive Rate (FPR) and Mean Absolute error (MAP) to hold that SCAK-NN is better than Decision Trees and Naïve Bayes.