Comparison of Multinomial Naïve Bayes Classifier, Support Vector Machine, and Recurrent Neural Network to Classify Email Spams

摘要

Information is one of the basic needs of the present. Many services are used as a means to send information messages, one of which is email. With the various conveniences offered by email, causing positive and negative impacts. One of the negative impacts is that an abused email becomes a spam email that has the potential to be detrimental and harm to the recipient. This is the basis for researching to determine the most accurate estimation of spam in emails by comparing the Multinomial Naïve Bayes Classifier algorithm(MNBC), Support Vector Machine algorithm(SVM), and Recurrent Neural Network(RNN) algorithm. The results of the evaluation are presented using the Classification Report to determine the results of accuracy, precision, memory, f1 score of each algorithm used. In this research, the algorithm that produces the greatest accuracy value in spam classification on email is the Support Vector Machine algorithm where the accuracy value of this algorithm is 96%, the precision value is 0.92, recall value is 0.96, and f1-score is 0.94.