摘要
由于提取的网络入侵特征主要以离散形式分布,波动性较大,会对后续检测准确性造成影响。为此,提出基于改进逻辑回归算法的电力监控系统网络入侵自动化检测系统。依据系统的基本要求与实现功能,设计四层物理架构,并对系统的硬件部分进行设计,包括网络数据探测器和核心交换机,基于此,使用one-hot编码方法对离散值进行编码处理,将连续数据转换为统一尺度的平稳序列,由此提取入侵特征,利用逻辑回归算法建立入侵检测模型,并引入梯度权重因子优化模型参数,将入侵特征作为模型输入,数据的异常分数作为输出,进而将待测流量数据的异常分数与基础阈值相比较,以此实现网络入侵检测。系统性能测试表明,利用设计的系统进行网络入侵检测,得到的检测漏报概率始终控制在0.2%以下,检测准确度较高。
Due to the fact that the extracted network intrusion features are mainly distributed in a discrete form with high volatility,it will affect the accuracy of subsequent detection.Therefore,a network intrusion automation detection system for power monitoring systems based on improved logistic regression algorithm is proposed.Based on the basic requirements and implementation functions of the system,a four layer physical architecture is designed,and the hardware part of the system is designed,including network data detectors and core switches.Based on this,the one hot encoding method is used to encode discrete values,convert continuous data into a uniform scale stationary sequence,extract intrusion features,establish an intrusion detection model using logistic regression algorithm,and introduce gradient weight factor optimization model parameters.The intrusion features are used as model inputs,and the abnormal score of the data is used as output.Then,the abnormal score of the tested traffic data is compared with the basic threshold to achieve network intrusion detection.The system performance test shows that using the designed system for network intrusion detection,the detection miss probability is always controlled below 0.2%,and the detection accuracy is high.
作者
曾体健
谢志奇
张玉吉
ZENG Tijian;XIE Zhiqi;ZHANG Yuji(Guizhou Wujiang Hydropower Development Co.,Ltd.,Guiyang 550002,China)
出处
《自动化与仪器仪表》
2025年第3期251-254,共4页
Automation & Instrumentation
基金
乌江公司水电电力监控系统生产控制大区网络安全建设项目技改项目编码(JG0120210125)。
关键词
改进逻辑回归算法
电力监控系统
入侵检测
网络数据探测器
检测模型
improve logistic regression algorithm
power monitoring system
intrusion detection
network data detector
detection model
