基于残差改进的灾变灰预测在电力行业网络安全预测中的应用

发布时间：2018-07-10 13:46

本文选题：灾变灰预测 + 残差改进　；参考：《河南大学》2014年硕士论文

【摘要】：针对电力行业内传统网络安全预测无法全面反映系统整体状况,预测精度不高的缺点,提出了一种网络安全预测方法。首先,对企业内的网络安全事件进行分析,运用层次分析法(AHP)构建网络安全的指标体系,并对样本数据进行过滤与处理,考虑异常值的分布序列结构:然后使用灾变灰预测进行模型的建模；最后对预测结果使用神经网络方法进行改进,从而达到提高预测精度的目的。通过仿真实验,表明基于残差改进的灾变灰预测方法的可行性和有效性。本文主要成果如下： 1对企业网络安全相关数据进行定期地采集,并进行分析整理,使用主成分分析法(PCA)提取出包含信息量较大的因素构成评价指标,将此作为企业网络安全预测的基础； 2对过滤后的信息进行处理,运用层次分析法(AHP)得出指标权重,构建网络安全评价的指标体系,为日后企业网络安全工作的开展提供依据； 3在样本数据分析的基础上,通过对神经网络方法及灰色模型方法的综合比较应用,提出一种适用于行业现状的预测模型,并进行数据处理,作为预测模型的输入,采用灾变灰预测模型对时间序列进行预测,并将结果代入神经网络模型进行修正,得到准确的网络安全预测值。
[Abstract]:Aiming at the shortcoming that the traditional network security prediction in the electric power industry can not reflect the whole system situation and the prediction accuracy is not high, a network security prediction method is proposed. First of all, the network security events in enterprises are analyzed, the index system of network security is constructed by AHP, and the sample data is filtered and processed. Considering the distribution sequence structure of outliers, the model is modeled with catastrophe grey prediction, and the prediction result is improved by neural network method to improve the prediction accuracy. The simulation results show the feasibility and effectiveness of the disaster grey prediction method based on residual error. The main achievements of this paper are as follows: 1 collecting and analyzing the related data of enterprise network security regularly, using principal component analysis (PCA) to extract the factors which contain a large amount of information to constitute the evaluation index. This is regarded as the basis of enterprise network security prediction. 2 the filtered information is processed and the index weight is obtained by using AHP to construct the index system of network security evaluation. It provides the basis for the work of enterprise network security in the future. 3 on the basis of sample data analysis, the comprehensive comparison between neural network method and grey model method is carried out. This paper presents a forecasting model suitable for the present situation of the industry, and carries on the data processing. As the input of the prediction model, the catastrophic grey prediction model is used to predict the time series, and the result is modified by the neural network model. Get accurate network security prediction value.
【学位授予单位】：河南大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.08

【参考文献】