Paper Details

Name: Gurpreet Kaur, Arvind Kumar & Kamal Malik
Country: India

Year: 2025
Volume: Volume-12, Issue-2 (July-December)
Page Number: 257-271
DOI: https://doi.org/10.5281/zenodo.17735137

ABSTRACT
Due to rapid advancements in network technologies, the digital landscape has become quite
complex. Accurate prediction of future traffic patterns have become crucial and critical for
optimal resource allocation and effective network management. So, in response to the challenge,
this study introduces a comprehensive deep learning framework that is designed to enhance the
accuracy of network traffic classification. Through the integration of CNN-based spatial feature
extraction and LSTM-driven temporal modeling, the hybrid design enhances classification
accuracy, compared to the conventional methods. The proposed model is applied to WSN-DS
dataset that achieved an impressive classification accuracy of 98%, verified using standard
performance evaluation metrics. A comprehensive comparative analysis against state-of-the-art
techniques further demonstrated its enhanced effectiveness across multiple performance
indicators. These outcomes highlight the framework‟s robustness, scalability, and practical
applicability, establishing it as a promising solution for real-time network monitoring and
intelligent traffic analysis in increasingly complex network environments.
KEYWORDS:
Deep learning, network traffic prediction, convolutional neural networks, long-short term memory,
wsn-ds dataset