摘要
针对目前使用的定量评估技术、Markov评估技术容易受到未授权第三方入侵影响,导致评估结果不精准的问题,提出了信息网络数据低延迟查询安全性回归预测模型方法。采用了低延迟复杂动态网络稳定性设计方案,保证网络的稳定性,避免了未授权第三方入侵干扰。在稳定网络环境下,对回归预测模型进行结构设计,将其划分为结果显示、核心预测、数据库三大模块。利用支持向量机训练样本集,监控网络态势值,构造安全态势样本。基于训练结果建立回归预测模型,并据此设计预测流程。建立了局部非线性模型,剔除过多的邻近点,使预测误差达到最小。在此基础上,发布网络安全预警信息。由实验结果可知,该模型最高预测精度为0.93,且安全态势值与实际值一致,为信息网络数据安全查询提供技术支持。
Aiming at the problem that the quantitative evaluation technology and Markov evaluation technology used at present are easily affected by the invasion of unauthorized third parties,which leads to inaccurate evaluation results,this paper proposes a regression prediction model method of information network data low-delay query security. A low-delay complex dynamic network stability design scheme is adopted to ensure the stability of the network and avoid intrusion interference by unauthorized third parties. Under the stable network environment,the structure of regression prediction model is designed and divided into three modules: result display,core prediction and database. Support vector machine is used to train sample sets,and the security situation samples are constructed by monitoring the network situation values. Based on the training results,a regression prediction model is established,and the prediction process is designed accordingly. A local nonlinear model is established to eliminate too many adjacent points and minimize the prediction error. On this basis, the network security warning information is released. The experimental results show that the maximum prediction accuracy of the model is 0.93,and the security situation value is consistent with the actual value,which provides technical support for information network data security query.
作者
马怡璇
李浩升
黄强
鲁学仲
王庆鹏
MA Yixuan;LI Haosheng;HUANG Qiang;LU Xuezhong;WANG Qingpeng(State Grid Xinjiang Electric Power Co.,Ltd.,Information and Communication Company,Urumqi 830000,China)
出处
《电子设计工程》
2023年第3期155-158,163,共5页
Electronic Design Engineering
关键词
网络低延迟
信息网络数据查询
安全性
回归预测模型
network low latency
information network data query
security
regression prediction model
