Performance Analysis of Anti-Phishing Tools and Study of Classification Data Mining Algorithms for a Novel Anti-Phishing System

The term Phishing is a kind of spoofing website which is used for stealing sensitive and important information of the web user such as online banking passwords, credit card information and user's password etc. In the phishing attack, the attacker generates the warning message to the user about the security issues, ask for confidential information through phishing emails, ask to update the user's account information etc. Several experimental design considerations have been proposed earlier to countermeasure the phishing attack. The earlier systems are not giving more than 90 percentage successful results. In some cases, the system tool gives only 50-60 percentage successful result. In this paper, a novel algorithm is developed to check the performance of the anti-phishing system and compared the received data set with the data set of existing anti-phishing tools. The performance evaluation of novel anti-phishing system is studied with four different classification data mining algorithms which are Class Imbalance Problem (CIP), Rule based Classifier (Sequential Covering Algorithm (SCA)), Nearest Neighbour Classification (NNC), Bayesian Classifier (BC) on the data set of phishing and legitimate websites. The proposed system shows less error rate and better performance as compared to other existing system tools.