Comparison And Simulation Of Several Kinds Of Estimation Methods For Copula And Its Related Functions

The Copula function is essentially a "join function",which is used to connect the joint distribution function of the multidimensional random variable with the one-dimensional marginal distribution function.The main function of the function of the extreme Copula function is to describe the correlation structure between multiple variables.In this paper,the estimation method of the Copula function and the estimation method of the relative function of the extremum Copula function are compared,and the kernel estimation method of the correlation function is introduced and analyzed in detail under the premise of non-parametric kernel estimation.I use the Copula function's estimation method to empirically analyze the correlation between the Shenzhen Stock Exchange and the daily logarithmic returns of the Shanghai Stock Exchange.At the end of the time,the method of the function of the Copula function is evaluated by a random simulation of the function of the function of the Copula function,and the method of estimating the method and the estimate of the method and the method of nuclear estimation is to be able to analyze it in a different way that the data is pure and unknown.The main structure of this article is as follows:1.The first part of this paper presents the research status of two function estimation methods first.In the second part,the basic concepts of the Copula function,such as the theory,nature and correlation coefficient of the Copula function are briefly described and introduced,and the commonly used model of Copula function is listed.2.The third chapter first introduces copulas connect function kernel density estimation method,and then through the comparison of kernel density estimation to get financial data copulas connect function and the commonly used copulas connect function model to predict the actual effect of VaR.3.The fourth chapter describes the correlation function of the binary extreme value Copula function,as well as the nature of the correlation function and some common models.And then introduces some bivariate extremal copulas connect function of correlation function estimation way,including HT-estimates,the CFG-estimation and OLS-estimation model.And the correlation function of the model,simulation analysis through the results can be concluded that OLS-estimated to be three kinds of estimation method to estimate the effect optimal estimation method.4.The fifth chapter firstly introduces the related function of extremum copulas connect function estimation methods including: three common nuclear nuclear estimates,P-C estimation,N-W estimation and G-M kernel estimation method.In the case of kernel function is continuous function of N-W estimation and P-C kernel estimate stochastic simulation analysis.Then select respectively in the case of pure and fuzzy data of two different cases of N-W kernel estimation method and OLS-estimation law stochastic simulation comparison,through the simulation results,by contrast,known in the condition of known data pure distribution of N-W nuclear estimation model and OLS-estimation model can be selected.However,the estimation effect of N-W kernel estimation model is better than OLS-estimation method in the case of unknown data.