Enhancing Spam Email Classification using Multilayer Perceptron: Performance Analysis and Comparative Evaluation
Ghizlane Khababa, Nor Hazlyna Harun, Sadik Bessou, Chu Sin Ning, Ooi Poa Ning · 2023
With the rapid increase in spam email volume and the associated productivity and security risks, effective spam email classification mechanisms are crucial in ensuring smooth communication systems. After several studies on spam email classification, the outcome demonstrated that Multilayer Perceptron (MLP) outperforms SVM and Decision Trees, making it the chosen model for further research in this paper, as it presented a study on spam email classification using a multilayer perceptron (MLP) model. The proposed MLP-based approach demonstrates strong performance in distinguishing between legitimate and spam emails. A comprehensive evaluation is conducted using a dataset of 5,172 emails, resulting in an accuracy rate of 94.40% on the testing set. The analysis of the results reveals interesting patterns and trends in the data, providing valuable insights into the effectiveness of the MLP model. Furthermore, a comparison with some state-of-the-art approaches showcases the superiority of the MLP model in terms of accuracy, F1 score, precision, recall, and MSE. The results of this study aid in the creation of reliable and precise spam email classification systems, increasing productivity and defending communication channels against spam threats.