基于数据增强和集成学习的网络流量检测  

Network Traffic Detection Based on Data Augmentation and Ensemble Learning

作　　者：吴苏亚 丁要军 WU Suya;DING Yaojun(School of Cyberspace Security,Gansu University of Political Science and Law,Lanzhou Gansu 730000,China)

机构地区：[1]甘肃政法大学网络空间安全学院,甘肃兰州730000

出　　处：《通信技术》2023年第11期1275-1282,共8页Communications Technology

摘　　要：针对网络流量检测中普遍存在的恶意流量样本不平衡问题,提出了一种基于数据增强和集成学习的分类方法。该方法采用K-means+SMOTE数据增强来平衡不同类别的数据样本,然后利用集成学习模型提高分类模型的泛化能力。在两个公开数据集上的实验结果表明,使用XGBoost进行二元和多类分类时,准确率分别达到99.1%和97.19%。与CNN、随机森林和LightGBM等模型相比,所提方法始终表现出显著性能优势。To address the common issue of imbalanced malicious traffic samples in network traffic detection,this paper proposes a classification method based on data augmentation and ensemble learning.The approach first employs K-means+SMOTE data augmentation to balance data samples of different categories,and then uses an ensemble learning model to enhance the generalization capability of the classification model.Experimental results on two publicly available datasets indicate that when using XGBoost for binary and multi-class classification,the accuracy reaches 99.1%and 97.19%,respectively.Compared to models such as CNN,Random Forest,and LightGBM,the approach proposed in this paper consistently shows significant performance advantages.

关 键 词：网络流量检测 数据增强 集成学习 流量识别 

分 类 号：TP393.0[自动化与计算机技术—计算机应用技术]