News

 Statistics

Number of Journals34
Number of Issues1,306
Number of Articles9,427
Article View9,188,328
PDF Download5,620,752
Majidian, S., TaghipourEivazi, S., Arasteh, B., Ghaffari, A. (2025). Computer Networks traffic classification model based on DBScan clustering and gamma classification. Electronic and Cyber Defense, (), -.
Seyede Zohreh Majidian; Shiva TaghipourEivazi; Bahman Arasteh; Ali Ghaffari. "Computer Networks traffic classification model based on DBScan clustering and gamma classification". Electronic and Cyber Defense, , , 2025, -.
Majidian, S., TaghipourEivazi, S., Arasteh, B., Ghaffari, A. (2025). 'Computer Networks traffic classification model based on DBScan clustering and gamma classification', Electronic and Cyber Defense, (), pp. -.
Majidian, S., TaghipourEivazi, S., Arasteh, B., Ghaffari, A. Computer Networks traffic classification model based on DBScan clustering and gamma classification. Electronic and Cyber Defense, 2025; (): -.

Computer Networks traffic classification model based on DBScan clustering and gamma classification

پدافند الکترونیکی و سایبری
Articles in Press, Accepted Manuscript, Available Online from 17 June 2025
Document Type: Original Article
Authors
Seyede Zohreh Majidian1; Shiva TaghipourEivazi* 2; Bahman Arasteh3; Ali Ghaffari4
1PhD student. Department of Computer Engineering, Aras international Branch, Islamic Azad University, Tabriz, Iran
2assistant professor. Department of Computer Engineering, Tabriz Branch, Islamic Azad University, Tabriz, Iran
3Associate Professor. Department of Computer Engineering, Tabriz Branch, Islamic Azad University, Tabriz, Iran
4Associate Professor . Department of Computer Engineering, Tabriz Branch, Islamic Azad University, Tabriz, Iran
Receive Date: 06 February 2025,  Revise Date: 07 April 2025,  Accept Date: 09 June 2025 
Abstract
Traffic classification is one of the most important network monitoring processes that has wide applications in the fields of security, quality of service, and network management. With the increasing complexity and variety of network traffic, new challenges arise, including the lack of labeled training data. In order to solve this challenge, in this paper, a traffic classification mechanism is presented by combining unsupervised and semi-supervised machine learning algorithms. This mechanism uses a limited set of labeled training data to improve classification accuracy. The proposed method describes each traffic flow as a feature vector that contains the statistical characteristics of that flow. The number of features generated for each sample is reduced using principal component analysis. DBScan clustering is used to determine the correct traffic type for each untagged traffic stream. Finally, the gamma classifier model is used to separate the new traffic flows. The efficiency of the proposed method has been evaluated using real data sets. The results show that the proposed method is able to classify traffic flows with an average accuracy of 95.12%, which shows at least 7.03% improvement over previous approaches.
Keywords
Traffic classification; Machine learning; DBScan clustering; Gamma classification
Main Subjects
Information security, encryption, encryption, protocols and standards
References
Statistics
Article View: 38