在网络恶意流量识别任务中,存在恶意流量样本数量与正常流量样本比例不平衡问题,从而导致训练出的机器学习模型泛化能力差,识别准确率低.为此,在网络流量图片化的基础上提出一种利用具有梯度惩罚项的条件Wasserstein生成对抗网络(CWGAN-GP)对少量数据类进行平衡的分类方法.该方法首先借助网络流量图片化方法将原始流量PCAP数据...
1.基于CWGAN‑GP星地链路信道建模的SCMA方法,其特征在于,所述方法包括: S1、构建卷积神经网络稀疏码多址接入SCMA模型,包括稀疏码多址接入SCMA的编码器、 星地链路信道和稀疏码多址接入SCMA的译码器; 所述星地链路信道采用生成对抗网络CWGAN‑GP实现,包括采用多个卷积神经网络单元 ...
基金项目:国家自然科学基金面上项目( 61472189 );赛尔网络下一代互联网技术创新项目( NGII20160105 )ThisworkwassupportedbytheNationalNaturalScienceFoundationofChina ( 61472189 ), CERNETInnovationProject ( NGII20160105 ) .通信作者:魏松杰( swei@njust.edu.cn )基于深度森林与 CWGANGGP 的移动应用网络行为分类与...
该方法首先借助网络流量图片化方法将原始流量PCAP数据按照流为单位进行切分,填充,映射到灰度图片中;然后使用CWGAN-GP方法实现数据集的平衡;最后,在公开数据集USTC-TFC2016和CICIDS2017上使用CNN模型对不平衡数据集和平衡后的数据集进行分类测试.实验结果表明,使用CWGAN-GP的平衡方法在精确度,召回率,F1这3个指标上均优...
并且其具有高精确率,高召回率,高F1-Score和低训练时间的特点.此外,为了解决应用样本数量有限且数据获取时间开销大等难题,还提出了一种使用CWGAN-GP的数据增强方法.与原始生成对抗网络相比,该模型训练更加稳定,仅需一次训练即可生成指定类别的数据.实验结果表明,在加入生成数据共同训练深度森林模型后,其分类准确率提高了...
基于CWGAN-GP 平衡化的网络恶意流量识别方法doi:10.12178/1001-0548.2022011丁要军王安宙Journal of the University of Electronic Science & Technology of China
This paper proposes a new data enhancement method combining the resampling and Conditional Wasserstein Generative Adversarial Networks-Gradient Penalty (CWGAN-GP), and uses the gray images-based Convolutional Neural Network (CNN) to realize the intelligent fault diagnosis of rolling bearings. First, the...
To this end, a novel data augmenting approach called GraphCWGAN-GP is proposed in this paper. The traffic data is first converted into grayscale images as the input for the proposed model. Then, the minority class data is augmented with our proposed model, which is built by...
CWGAN-GP models can generate data that resembles experimental data in metabolomics.Supplementing the minority class with generated samples improves model performance.Improvement in model performance increases with class separation.A set of criteria is proposed for the evaluation of quality of generated data...
CWGAN-GPNeural Computing and Applications - Traditional machine learning methods are difficult to obtain good performance in the classification of gene expression data due to its characteristics of high...doi:10.1007/s00521-022-07417-9Han, Fei...