An Analysis of Supervised Machine Learning Algorithms for Spam Email Detection
Tasnia Jahan Toma, Samia Binta Hassan, Mohammad Arifuzzaman · 2021 International Conference on Automation, Control and Mechatronics for Industry 4.0 (ACMI) · 2021
Email users are increasing at a high rate and a huge number of people’s privacy is getting risked by spam email and it also kills valuable time of people most often. Spam email can be malicious as well as it can be of commercial use as in for marketing which are not desirable to us. Hence, detecting and filtering spam emails from several emails is a must work to do. There are enormous machine learning (ML) algorithms and some of them can be used to detect and analyze spam and unwanted emails. In this paper, we use the supervised ML technique on an existing email classification dataset where we explore Naïve Bayes, Support Vector Machine, Random Forest Classifier. Along with observing the accuracy from these algorithms, we showed other performance metric like precision, recall and F1 score etc. We got a high rate of accuracy in each algorithm such as we got 98.8%, 97.6%, 91.5%, 97.8%, 98.5% accuracy in Multinomial Naïve Bayes, Bernoulli Naïve Bayes, Gaussian Naïve Bayes, Random forest classifier, Support vector machine (SVM) respectively.