Email Spam Filtering Using Machine Learning Based Xgboost Classifier Method

Abstract

E-mail spam, known as undesirable Bulk E-mail (UBE), junk mail, or undesirable commercial e-mail (UCE), is transferring undesirable e-mail information, usually with business data, in large amounts to a confused set of recipients. Spam is standard on the Internet because automated communications' transaction costs are lower than other alternative forms of communication. Many spam filters use various approaches to recognize the incoming message as spam, varying from white list/blacklist, Bayesian review, keyword matching, postage, mail header analysis, enactment, etc. Even though we are still involved in spam e-mails every day, this paper proposed an enhanced spam exposure design based on Extreme Gradient Boosting (XGBoost) model. It is studied for increased accuracy in spam detection. To the best of our experience, it has expected minor considerations spam e-mail detection difficulties.  We explore the proposed system's performance using a more comprehensive range of experimental metrics behind the accuracy, which has managed the existed investigations. The proposed algorithm comparing with existed classifiers of SVM, CNSA-FFO, Rotation forest, MLP, J48, and Naïve Bayes. The proposed model gets better accuracy with 95% when compared with previous classifiers