Variable Coefficient Std. Error t-Statistic Prob.  
   
AR(1) -0.862433 0.045570 -18.92528 0.0000
MA(1) 1.430487 0.314362 4.550445 0.0001
   
R-squared 0.493010     Mean dependent var  -0.000265
Adjusted R-squared 0.474903     S.D. dependent var  0.066488
S.E. of regression 0.048180     Akaike info criterion  -3.163421
Sum squared resid 0.064996     Schwarz criterion  -3.070008
Log likelihood 49.45131     Hannan-Quinn criter.  -3.133537
Durbin-Watson stat 1.528937   
   上面的6.0
===============================================   
下面的是3.1
Variable Coefficient Std. Error t-Statistic Prob.  
   
AR(1) -0.847454 0.009093 -93.20302 0.0000
MA(1) 1.274697 0.239579 5.320564 0.0000
   
R-squared 0.408639     Mean dependent var  -0.000265
Adjusted R-squared 0.387519     S.D. dependent var  0.066488
S.E. of regression 0.052034     Akaike info criterion  -3.009486
Sum squared resid 0.075812     Schwarz criterion  -2.916072
Log likelihood 47.14228     F-statistic  19.34838
Durbin-Watson stat 1.597869     Prob(F-statistic)  0.000143
   
Inverted AR Roots       -.85   
Inverted MA Roots       -1.27   
Estimated MA process is noninvertible   
   
求高人解答