非对称指数幂分布与正态分布VAR模型的模拟效果评价
|
摘要
文章把非对称指数幂分布(AEPD)标准化为标准非对称指数幂分布(SSAEPD),通过高斯连接函数(Gaussian Copula)将其应用到向量自回归模型(VAR)中,构造一个新的模型——VAR-GC-SSAEPD模型,并利用Matlab生成多元时间序列的模拟数据,分别拟合正态分布的向量自回归模型和VAR-GC-SSAEPD模型,然后采用赤池信息准则(AIC)、施瓦茨准则(SC)以及对模型真实参数估计的准确度来比较两个模型的优劣。最终得出VAR-GC-SSAEPD模型相比于VAR模型更优的结论。
This paper firstly normalizes the asymmetric exponential power distribution(AEPD) to be a standard asymmetric exponential power distribution(SSAEPD), and applies it to a vector auto-regression(VAR) model through Gaussian Copula to construct a new model-VAR-GC-SSAEPD model. Then the paper uses Matlab to generate simulation data of multivariate time series to fit the normally distributed vector autoregression model and VAR-GC-SSAEPD model respectively. The paper also uses AIC criterion value, SC criterion value and the accuracy of model real parameter estimation to compare the advantages and disadvantages of the two models. Finally, it is concluded that the VAR-GC-SSAEPD model is better than the VAR model.
This paper firstly normalizes the asymmetric exponential power distribution(AEPD) to be a standard asymmetric exponential power distribution(SSAEPD), and applies it to a vector auto-regression(VAR) model through Gaussian Copula to construct a new model-VAR-GC-SSAEPD model. Then the paper uses Matlab to generate simulation data of multivariate time series to fit the normally distributed vector autoregression model and VAR-GC-SSAEPD model respectively. The paper also uses AIC criterion value, SC criterion value and the accuracy of model real parameter estimation to compare the advantages and disadvantages of the two models. Finally, it is concluded that the VAR-GC-SSAEPD model is better than the VAR model.
引文
[1] Zhu D, Zin-Walsh V. Properties and Estimation of Asymmetric Exponential Power Distribution[J]. Journal of Econometrics, 2009, 148(1).
[2]Zhu D, Galbraith J W. Modeling and Forecasting Expected Shortfall With the Generalized Asymmetric Student-t and Asymmetric Exponential Power Distributions[J]. Journal of Emperical Finance, 2011,Vol.18(4).
[3]Li L, Gan Q, Zhuo Z, et al. Testing the CAPM Theory Based on a New Model for Fama-French 25 Portfolio Returns[J]. Theoretical Economics Letters, 2014, 4(8).
[4]靳慧杰.基于非对称指数幂分布的资本资产定价模型研究[D].天津:南开大学硕士学位论文,2011.
[5]李孝华,宋敏.基于AEPD分布和ALD分布的VaR模型[J].数量经济技术经济研究, 2013,(1).
[6]邸昊,李柳玲. CAPM的新模型研究——基于美国银行信用违约互换数据的研究[J].河北民族师范学院学报, 2015,(2).
[7]沈悦,李善燊,马续涛. VAR宏观计量经济模型的演变与最新发展——基于2011年诺贝尔经济学奖得主Smis研究成果的拓展脉络[J].数量经济技术经济研究, 2012,(10).
(1)AIC=-2(ln(L)/T)+2m/T,SC=-2(ln(L)/T)+(mln(T))/T,m=N(l+KN),N为变量个数,k为滞后期,l为常数项个数,L为极大似然函数值。
[2]Zhu D, Galbraith J W. Modeling and Forecasting Expected Shortfall With the Generalized Asymmetric Student-t and Asymmetric Exponential Power Distributions[J]. Journal of Emperical Finance, 2011,Vol.18(4).
[3]Li L, Gan Q, Zhuo Z, et al. Testing the CAPM Theory Based on a New Model for Fama-French 25 Portfolio Returns[J]. Theoretical Economics Letters, 2014, 4(8).
[4]靳慧杰.基于非对称指数幂分布的资本资产定价模型研究[D].天津:南开大学硕士学位论文,2011.
[5]李孝华,宋敏.基于AEPD分布和ALD分布的VaR模型[J].数量经济技术经济研究, 2013,(1).
[6]邸昊,李柳玲. CAPM的新模型研究——基于美国银行信用违约互换数据的研究[J].河北民族师范学院学报, 2015,(2).
[7]沈悦,李善燊,马续涛. VAR宏观计量经济模型的演变与最新发展——基于2011年诺贝尔经济学奖得主Smis研究成果的拓展脉络[J].数量经济技术经济研究, 2012,(10).
(1)AIC=-2(ln(L)/T)+2m/T,SC=-2(ln(L)/T)+(mln(T))/T,m=N(l+KN),N为变量个数,k为滞后期,l为常数项个数,L为极大似然函数值。