Network traffic anomaly detection method based on multi-scale convolution and channel attention mechanism

Considering the problems of traditional detection methods limited by weak model representation capabilities and vulnerability to data class imbalance, a network traffic anomaly detection method integrating multi-scale convolution and a channel attention mechanism was proposed. Firstly, a pyramid con...

ver descrição completa

Detalhes bibliográficos

Principais autores:	Fu Yu, Wang Yujue, Yu Yihan, Liu Taotao, An Yishuai
Formato:	Artigo
Idioma:	Chinês
Publicado em:	Editorial Department of Journal on Communications 2026-01-01
Colecção:	Tongxin xuebao
Assuntos:	network traffic anomaly detection multi-scale convolution attention mechanism balanced loss function
Acesso em linha:	http://www.joconline.com.cn/zh/article/doi/10.11959/j.issn.1000-436x.2026010/
Tags:	Adicionar Tag Sem tags, seja o primeiro a adicionar uma tag!