Network Resource Risk Forecast Based On Copula Theory

With the rapid development of cloud computing,massive storage and new network technology,the security and maintenance of network is faced with many problems.Firstly,the scale of network increasingly large and network bandwidth is growing rapidly,the existing network monitoring facilities lack of enough monitoring methods and extensibility to deal with this change.Secondly,the network applications are so plentiful and all kinds of network threats hidden in the traffic so that the network resources are strained,the network performance is declining and collapse,we need to predict the risk as soon as possible and assess appropriately so as to protect the network resources.Therefore,it is important to predict and assess the risk of network resources timely and effective.This paper studies on the prediction and assessment of network resources risk.The previous prediction methods of the network resources risk are mostly aimed at one of it,so this paper selected several network resource indexes,proposed a forecasting method based on vine-Copula-GARCH model and assessed the risk of network by Bayesian networks.Firstly,this paper studied the prediction method of network resources risk based on the theory of the Copula.According to several network resources indexes,we collected the relevant data respectively.After normalizing treatment,calculating the edge of their distribution function and density function combined with the GARCH model.Then,combining the theory of Copula and the vine structure,we confirmed the logical structure of the vine through the correlation coefficient.Building the model which is uesd to calculate the multivariate joint distribution function by the pair-copula decompose model and calculating the VaR of network resources by the monte carlo simulation.Secondly,this paper discussed the assess method of the network resources by Bayesian Network on the basis of the network resources risk prediction.According to the network topology and the analysis of the collected network resources data,the method calculated the priorprobability of each index and obtained the posterior probability through the BNT toolbox.At last,obtaining the sort of degree that the risk of network has been affected by every networkresource and the level of risk.Finally,the experimental results show that the prediction method of the risk of network resource based on the theory of the Copula,simplified the calculation in accordance with the advantage that the Copula separated the dependent structure of variables and the marginal distribution,the multivariate joint distribution is decomposed into the binary Copula function.According to the result that compared the predicted value with the given threshold,assessing the risk of network resources by Bayesian Network.This method has good effect for assessment and gets the desired purpose.