The next section contains the coefficient estimates, standard errors, t-statistics and corresponding
p-value. (It is worth pointing out that the reported ARMA coefficients use a different
sign convention than those in Sowell so that the ARMA coefficients all have the opposite
sign).
Notice that since we estimated the model using ML, EViews displays the estimate of the
error variance as one of the estimated coefficients将误差方差作为估计的系数. You should be aware that the EViews
reported p-value for SIGMASQ is for the two-sided test双侧检验??, despite the fact that SIGMASQ must
be non-negative. (If desired, you may use the reported coefficient, standard error, and the
@CTDIST function to compute the appropriate one-sided p-value.)
我在eviews9的说明中找到这样一段,但看的不是太明白
|