引用本文:王娜,张鑫海,常娅明.基于变分客观模糊辨识的态势预测[J].控制理论与应用,2025,42(9):1789~1797.[点击复制]
WANG Na,ZHANG Xin-hai,CHANG Ya-ming.Situation prediction based on variational objective fuzzy identification[J].Control Theory & Applications,2025,42(9):1789~1797.[点击复制]
基于变分客观模糊辨识的态势预测
Situation prediction based on variational objective fuzzy identification
摘要点击 1904  全文点击 146  投稿时间：2023-11-16  修订日期：2025-05-16
查看全文  查看/发表评论  下载PDF阅读器   HTML
DOI编号  10.7641/CTA.2024.30744
  2025,42(9):1789-1797
中文关键词  态势预测  T-S模型  模糊辨识  变分模态分解  动态时间规整  模糊聚类
英文关键词  situation prediction  T-S model  fuzzy identification  variational mode decomposition  dynamic time warp ing  fuzzy clustering
基金项目  天津市自然科学基金重点项目(23JCZDJC01140)资助.
作者单位E-mail
王娜* 天津工业大学控制科学与工程学院 wangna@tiangong.edu.cn 
张鑫海 天津工业大学控制科学与工程学院  
常娅明 天津工业大学控制科学与工程学院  
中文摘要
      针对网络安全态势数据具有较强的非平稳性和随机性,为提高网络安全态势预测的精度,提出一种基于变 分客观模糊辨识的态势预测方法.首先,引入变分模态分解,并与动态时间规整方法相结合,来对原始态势数据集 进行分解和重构,以提高该数据集的平稳性并减少分解后的模态个数,从而降低后续模型预测的误差和训练成本; 然后,利用偏自相关分析确定T-S模型的输入变量,并结合客观聚类分析和模糊c均值聚类,来直接获取紧凑而准确 的T-S模型结构,从而确保所建预测模型的精确性;最后,通过标准网络入侵检测数据集NSL-KDD的仿真来验证所 提方法的有效性.
英文摘要
      In view of the strong non-stationarity and randomness existing in the network security situation data, to improve the accuracy of network security situation prediction, a method via variational objective fuzzy identification is proposed. Firstly, the variational mode decomposition method is introduced and combined with the dynamic time warping method. Thus the original situation data set is decomposed and reconstructed. As a result, the stability of the proposed data set is increased and the number of decomposed modes is reduced. Simultaneously, the error and training cost of the subsequent model predictions is decreased; Secondly, the partial autocorrelation analysis method is used to determine the input variables of the T-S model. Followingly, the inputs are afforded to the algorithms of objective cluster analysis and the fuzzy c-means. Therefore the compact and accurate structure of the T-S model is obtained. As a result, the accuracy of the constructed T-S model is guaranteed. Finally, the validity of the presented method was tested using the NSL-KDD benchmark dataset of network intrusion detection simulation.