摘要
提出一种基于支持向量机和粒子群算法的网络态势复合预测模型。模型使用滑动窗口方法将各原始离散时间监测点的安全态势值构造成部分线性相关的连续时间序列,以其作为安全态势数据样本集对支持向量机加以训练,生成预测模型。在支持向量机训练过程中,利用粒子群算法搜寻支持向量机的最优训练参数,以降低支持向量机参数选择的盲目性,提高预测精度。最后通过基于大量电力企业信息网络现场安全监测数据的实验,验证了复合预测模型的有效性。
A security situation prediction model for information network based on support vector machine （SVM） and particle swarm optimization （PSO） is proposed. By use of sliding window, in the proposed model a continuous time series that is partially linearly dependent is constructed by security situation values sampled from original discrete time monitoring points, and taking the time series as the sample set of security situation data the SVM is trained to generate a prediction model. During the training of SVM, the PSO algorithm is used to search for the optimal training parameters of SVM to reduce the blindness in the selection of SVM parameters and improve precision of prediction. Through the experiments based on on-site installation and monitoring data of a lot of power enterprise information networks, the effectiveness of the proposed security situation prediction model is verified.
出处
《电网技术》
EI
CSCD
北大核心
2011年第4期176-182,共7页
Power System Technology
基金
国家电网公司科技项目(B11-09-109)
关键词
信息网络安全态势
回归预测
支持向量机
粒子群算法
时间序列
security situation of information network
regression prediction
support vector machine
particle swarm optimization
time series
