Hybrid Feature Selection and Ensemble Learning Method for Spam Email Classification

Abstract

The data mining techniques produce good work in many domains. The spam emails are becoming a serious dilemma and an important matter to have different solutions, and enhanced methods and algorithms. Using Ensemble methods which are well-established classifiers. In this paper data mining techniques used to classify spam email using the UCI spam base dataset. The results achieved by the machine learning tools and techniques, and the Ensemble learning methods, after applying feature selection methods on the data set; which gave better result, and better classification accuracy. For the evaluation method used the cross-validation for testing and training option, and the confusion matrix to show the accuracy and the performance result of the chosen classifiers; which are Naïve Bayes, decision tree, ensemble boosting and ensemble hybrid boosting classifiers.