摘要
针对传统电网信息系统安全风险预测模型没有分解信号时频特征,导致存在预测准确性能低、预测时间长的问题,提出基于机器学习算法的电网信息系统安全风险预测模型。在电网信息系统中安装数据采集装置,采集电网信息系统运行工况信息,利用自相关匹配检测方法,实现对采集工况信号时频特征的分解,基于交互式学习技术,完成采集运行数据的聚类分析,设置系统安全风险预测指标,并利用机器学习算法计算对应指标的权值,从系统网络攻击预测、用户行为预测和系统硬件设备故障预测3个方面,得出系统安全风险的预测值。实验结果表明:研究模型的预测准确率更高,预测时间更短,有效性更好。
In view of the problem of low prediction accuracy and long prediction time caused by the failure of the traditional network information system security risk prediction model to decompose the time-frequency characteristics of signals,a network information system security risk prediction model based on machine learning algorithm is proposed.Data acquisition device in the network information system,information acquisition power grid information system running,the use of the correlation matching method,realize the working condition of acquisition signal time-frequency characteristics of decomposition,based on the interactive learning technology,complete running data clustering analysis,set up the system security risk predictors,and machine learning algorithm is used to calculate the corresponding indicators weight,from network attack prediction system,user behavior prediction and three aspects of system hardware failure prediction system security risk prediction.The experimental results show that the prediction accuracy is higher,the prediction time is shorter and the prediction efficiency is better.
作者
陆冰芳
张希翔
LU Bing-fang;ZHANG Xi-xiang(Information Center of Guangxi Power Grid Co.,Ltd,Nanning 530023,China)
出处
《电子设计工程》
2020年第13期128-132,共5页
Electronic Design Engineering
基金
广西科技厅项目(2016071)。
关键词
机器学习
电网信息系统
系统安全
风险预测
模型构建
machine learning
grid information system
system security
risk prediction
model building