基于PCA-2KPCA-SVM的pod入侵高精度检测方法
发布时间：2019-05-19 06:52
【摘要】：为精确识别具体的攻击行为,提高入侵行为的识别率,提出一种基于PCA-2KPCA-SVM的pod入侵高精度检测方法。根据样本的不同分布特点选择不同的PCA方法,进行特征抽取和降维预处理;对于所有入侵行为样本,将pod入侵样本同其它入侵行为样本及正常样本做一对一分组;所有小组均使用PCA-PSO-SVM方法训练,将训练效果不佳的小组使用2KPCA-SVM方法训练,根据每个小组的训练方法对测试样本进行检测。实验结果表明,该方法可以精确识别pod入侵行为,可推广到对其它入侵行为的高精度检测。
[Abstract]:In order to accurately identify specific attack behavior and improve the recognition rate of intrusion behavior, a high precision detection method of pod intrusion based on PCA-2KPCA-SVM is proposed. According to the different distribution characteristics of the samples, different PCA methods are selected for feature extraction and dimension reduction preprocessing, and for all intrusion samples, the pod intrusion samples are grouped one-to-one with other intrusion samples and normal samples. All groups were trained by PCA-PSO-SVM method, and the groups with poor training effect were trained by 2KPCA-SVM method, and the test samples were tested according to the training methods of each group. The experimental results show that this method can accurately identify pod intrusion behavior and can be extended to high precision detection of other intrusion behaviors.
【作者单位】： 无锡职业技术学院物联网技术学院;江南大学物联网工程学院;
【基金】：江苏省产学研联创基金项目(BY2013015-40)
【分类号】：TP393.08
，
本文编号：2480491
[Abstract]:In order to accurately identify specific attack behavior and improve the recognition rate of intrusion behavior, a high precision detection method of pod intrusion based on PCA-2KPCA-SVM is proposed. According to the different distribution characteristics of the samples, different PCA methods are selected for feature extraction and dimension reduction preprocessing, and for all intrusion samples, the pod intrusion samples are grouped one-to-one with other intrusion samples and normal samples. All groups were trained by PCA-PSO-SVM method, and the groups with poor training effect were trained by 2KPCA-SVM method, and the test samples were tested according to the training methods of each group. The experimental results show that this method can accurately identify pod intrusion behavior and can be extended to high precision detection of other intrusion behaviors.
【作者单位】： 无锡职业技术学院物联网技术学院;江南大学物联网工程学院;
【基金】：江苏省产学研联创基金项目(BY2013015-40)
【分类号】：TP393.08
，
本文编号：2480491
本文链接：https://www.wllwen.com/guanlilunwen/ydhl/2480491.html
