AUTHOR=Liu Chao , Liu Chunxiang , Liu Changrong 

TITLE=An abnormal traffic detection method for chain information management system network based on convolutional neural network

JOURNAL=Frontiers in Physics

VOLUME=Volume 13 - 2025

YEAR=2025

URL=https://www.frontiersin.org/journals/physics/articles/10.3389/fphy.2025.1592975

DOI=10.3389/fphy.2025.1592975

ISSN=2296-424X

ABSTRACT=Chain information management system is widely used, providing convenience for the operation and management of enterprises. However, the problem of abnormal network traffic becomes increasingly prominent currently. Therefore, this paper proposes a convolutional neural network based on attention mechanism and autoencoder improvement, namely, CBAM-AE-CRF. CBAM AE-CRF integrates the convolutional block attention module (CBAM) into convolutional neural network to enhance the model’s ability to learn anomalous features in network traffic. CBAM improves the detection accuracy of abnormal traffic in chain information management system by adaptively adjusting channel attention. At the same time, the Autoencoder module (AE) is also introduced into the model to automatically extract and reconstruct anomalous features from complex network traffic data. Finally, the conditional random field (CRF) determines the optimal label sequence based on the conditional probability distribution and applies the Viterbi algorithm to complete the sequence labeling of network traffic in chain information management system. Through extensive experimental verification, CBAM-AE-CRF can comprehensively understand the semantics of network traffic, accurately identify anomalies in network traffic of chain information management system, provide strong support for network traffic management.