Comparison of Machine Learning Algorithms' Performance in Network Traffic Classification

Authors

  • faifaa micaiel Tishreen University
  • Jamal Khalifa Tishreen University
  • Mohannad Issa Ministry of Oil and Mineral Resources

Abstract

Maintaining network availability and improving performance is a primary objective of network management. With the massive growth in network size and traffic load, this task has become increasingly complex. One important area of research is network traffic classification, which offers significant benefits such as reducing traffic congestion and enhancing network management.

This study explores the application of various machine learning algorithms for network traffic classification into large flows and small flows. We implemented and evaluated multiple classifiers on real network traffic “Darknet Dataset”, including Support Vector Machine (SVM), Random Forest (RF), k-Nearest Neighbors (KNN), Logistic Regression (LR), Decision Tree (DT), Gradient Boosting (GB), and Multilayer Perceptron (MLP). Each classifier was trained and tested on real network traffic data.

Our results indicate that the Random Forest (RF), Decision Tree (DT), Nearest Neighbor (KNN) classifier, Gradient Boosting (GB), and Multilayer Neural Network (MLP) classifiers achieved the highest accuracy, in classification. These results underscore the potential of the models. Machine learning helps in classifying network loads effectively.

Published

2024-11-03

How to Cite

1.
مكائيل ف, جمال خليفة, مهند عيسى. Comparison of Machine Learning Algorithms’ Performance in Network Traffic Classification. Tuj-eng [Internet]. 2024Nov.3 [cited 2024Dec.27];46(4):177-98. Available from: https://journal.tishreen.edu.sy/index.php/engscnc/article/view/17643

Most read articles by the same author(s)