Email spam detection using bagging and boosting of machine learning classifiers

The increase in the popularity, utility, and significance of electronic mails has also raised the exposure of spam emails. This paper endeavours to detect email spam by constructing an ensemble system using bagging and boosting of machine learning techniques. The dataset used for the experimentation is Ling-Spam Corpus. The system detects spam email by bagging the machine learning-based multinomial Naïve Bayes (MNB) and J48 decision tree classifiers followed by the boosting technique of converting weak classifiers into strong by implementing the Adaboost algorithm. The experimentation includes three different experiments and the results attained are compared with each other. Experiment 1 employs the individual classifiers, experiment 2 ensembles the classifiers with bagging approach, and experiment 3 ensembles the classifiers by implementing the boosting approach for the email spam detection. The effectiveness of the ensemble methods is manifested by comparing the evaluated results with individual classifiers in terms of evaluation metrics.