Network traffic classification using Unsupervised Learning: a comparative analysis of clustering algorithms
Year: 2023
Type: preprint
Abstract: Network traffic classification plays a crucial role in various network management tasks, such as resource allocation, intrusion detection, and Quality Of Service (QOS) optimization. With the increasing complexity and volume of network traffic, the need for efficient and accurate classification algorithms has become paramount. Traffic classification plays a vital role in achieving QoS, but the task has become increasingly difficult due to evolving applications and growing network traffic. This article focuses on leveraging machine learning techniques, particularly unsupervised learning, for network traffic classification. The objective is to replicate the work of Aouedi et al. and compare the performance of various clustering algorithms: k-means, DBSCAN, bisecting k-means, and k-modes. A more refined experimental procedure is proposed, including data preparation, feature selection, hyperparameter analysis, and cluster analysis. This study contributes to the understanding of different clustering algorithms' performance for network traffic classification.
Source: HAL
Authors: Canever, Helena, Wang, Xihui
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W4390850477
