Using Machine Learning Algorithm to Enhance Phishing Emails Classification

Phishing emails are becoming a more dangerous problem in online bank truncation processing problem, as well as social networking sites such as Facebook, Twitter, and Instagram. Phishing is typically carried out by imitating an email or embedding content in the email body, which prompts consumers to submit their credentials. Training on the phishing strategy is not so much effective since users forget their training tricks and warning messages after a while. It is entirely dependent on the user's activity, which will be taken at a specific time in response to software warning messages while operating any URL. On the basis of the Spam base dataset, this paper improves phishing email classification using J48, Nave Bayes, and decision tree. J48 performs the best classification on spam basis, with a true positive rate of 97 percent and a false negative rate of 0.025 percent. Random forest work best on small dataset that is up to 5000 and number of feature are 34. However, when the dataset size is increased and the number of features is reduced, Nave Bayes works faster.