悬赏 50 个论坛币 未解决
各位大神
主要有两个问题:
1、用r做的GARCH拟合值与用eviews做出的GARCH拟合值不同是什么原因,一些默认方法不同?
2、代码的结果好像和想得到的结果不太一样,想要通过迭代得出最后一个GARCH拟合值组成的序列
部分代码:
library("timeDate")
library("timeSeries")
library("fBasics")
library("fGarch")
local_1_500=c()
for (i in 2:(length(e)-499)){
 fm=garchFit(~arma(1,0)+garch(1,1),data=e[i:(i+499)],cond.dist="norm")
 local_1_500[(i+499),]=summary(fm)$fitted[,1]   
}
得出部分结果:
--- END OF TRACE ---
Time to Estimate Parameters:
 Time difference of 0.1404002 secs
Title:
 GARCH Modelling 
Call:
 garchFit(formula = ~arma(1, 0) + garch(1, 1), data = e[i:(i + 
    499)], cond.dist = "norm") 
Mean and Variance Equation:
 data ~ arma(1, 0) + garch(1, 1)
<environment: 0x04d03350>
 [data = e[i:(i + 499)]]
Conditional Distribution:
 norm 
Coefficient(s):
        mu         ar1       omega      alpha1       beta1  
3.6900e-03  1.0000e+00  2.3229e-05  2.2057e-01  7.8901e-01  
Std. Errors:
 based on Hessian 
Error Analysis:
        Estimate  Std. Error  t value Pr(>|t|)    
mu     3.690e-03   1.212e-03    3.043  0.00234 ** 
ar1    1.000e+00   2.633e-03  379.834  < 2e-16 ***
omega  2.323e-05          NA       NA       NA    
alpha1 2.206e-01   3.569e-03   61.799  < 2e-16 ***
beta1  7.890e-01   9.204e-03   85.722  < 2e-16 ***
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 
Log Likelihood:
 956.6536    normalized:  1.913307 
Description:
 Mon Jul 18 15:33:09 2016 by user: Administrator 
Standardised Residuals Tests:
                                Statistic p-Value     
 Jarque-Bera Test   R    Chi^2  23.92063  6.392936e-06
 Shapiro-Wilk Test  R    W      0.989324  0.001058274 
 Ljung-Box Test     R    Q(10)  34.19756  0.0001709915
 Ljung-Box Test     R    Q(15)  41.11331  0.0003073162
 Ljung-Box Test     R    Q(20)  43.61511  0.001694057 
 Ljung-Box Test     R^2  Q(10)  9.093494  0.5232543   
 Ljung-Box Test     R^2  Q(15)  16.19706  0.3690775   
 Ljung-Box Test     R^2  Q(20)  21.46418  0.3702786   
 LM Arch Test       R    TR^2   10.47235  0.5745945   
Information Criterion Statistics:
      AIC       BIC       SIC      HQIC 
-3.806614 -3.764468 -3.806812 -3.790076 
Series Initialization:
 ARMA Model:                arma
 Formula Mean:              ~ arma(1, 0)
 GARCH Model:               garch
 Formula Variance:          ~ garch(1, 1)
 ARMA Order:                1 0
 Max ARMA Order:            1
 GARCH Order:               1 1
 Max GARCH Order:           1
 Maximum Order:             1
 Conditional Dist:          norm
 h.start:                   2
 llh.start:                 1
 Length of Series:          500
 Recursion Init:            mci
 Series Scale:              0.5386348
错误于arima(.series$x, order = c(u, 0, v), include.mean = include.mean) : 
  non-stationary AR part from CSS
此外: 共有44个警告 (用warnings()来显示)