Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{170673,
        author = {Adesh Bhosale and Aditya Kadam and Sejal Abraham and Prajwal Shankarshetty and Sangeeta Alagi and Milind Ankaleshwar},
        title = {Federated learning approach for multiple classes in network traffic classification},
        journal = {International Journal of Innovative Research in Technology},
        year = {2024},
        volume = {11},
        number = {7},
        pages = {1412-1416},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=170673},
        abstract = {Federated learning (FL) is an emerging paradigm in machine learning that allows decentralized model training across multiple devices or clients without the need to share raw data. This is especially valuable in scenarios where data privacy and security are paramount, such as in the case of network traffic classification. In the context of network traffic classification, FL enables multiple edge devices, such as routers, firewalls, or IoT devices, to collaboratively build a model capable of classifying traffic into multiple classes, such as normal traffic, Distributed Denial-of-Service (DDoS) attacks, malware, or other anomalous traffic. The key advantage of FL in this scenario is that it avoids the need to centralize sensitive traffic data, which can be privacy-sensitive and large in volume, especially when monitoring diverse network environments.

The federated learning process in network traffic classification begins with each client (e.g., an edge device or a network node) training a local model on its own network traffic data. The traffic data on each device may be highly heterogeneous, with some clients having more benign traffic while others might experience a higher volume of attack-related traffic. The local model is typically a neural network or another machine learning model that is trained to recognize patterns of normal and anomalous network behaviour Once the local model is trained on the device’s data, only model updates (such as gradients or weight adjustments) are sent to a central server, rather than the raw network traffic data itself.},
        keywords = {},
        month = {December},
        }