Email spam detection: a comparison of svm and naive bayes using bayesian optimization and grid search parameters

Abstract

Spam emails are still a big problem, crowding out inboxes and annoying email users everywhere. SVM and Naive Bayes are frequently used algorithms that have demonstrated excellent performance in performing text classification, including spam detection. The purpose of this study is to evaluate the overall performance of SVM and Naive Bayes in the context of detecting spam emails using default parameters. This research utilizes Bayesian Optimization and Grid Search Parameters for both SVM and Naive Bayes models to help maximize the performance of the constructed models. This study uses a spam email dataset that has 2 sample groups, namely spam and ham. Of the three parameter selection methods that have been tested on the SVM Algorithm, Bayesian Optimization is a parameter tuning method that has the most satisfying results in accuracy, precision, recall, and f1 scores respectively with values of 98.5642%, 99.4048%, 89.