MULTI STAGE PHISHING EMAIL CLASSIFICATION

Phishing emails risk increases progressively, which pose a real threat to users of computers, organizations and lead to significant financial losses. Fighting zero day phishing emails using content based server side classifiers is considered as the best method to detect such attacks. This technique which is based on machine learning algorithms is trained by the set of phishing email features and the statistical classifier is used on stream of email to detect the class of fresh email received. The false positive rate (FPR) and false negative rate (FNR) are critical factors for these classifiers and should be as small as possible to increase the overall accuracy of the classifiers. Using the ham and phishing data sets available, this article focuses on reduction of false positive rate (FPR), false negative rate (FNR), and increase the overall accuracy of the proposed classification system. The multi stage phishing email detection system shows very promising results compared with previous works in term of FPR, FNR, and accuracy.