Announcement

Collapse
No announcement yet.
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Q-test after ARIMA

    Dear All,

    Suppose I fit a simple ARIMA model:

    Code:
    arima Y, ar(1) ma(2)
    And if I type:

    Code:
    predict r1, resid
    and then I type:

    Code:
    wntestq r1, lags(12)
    Is it permissible to use the above command to check for the corresponding autocorrelation in residuals? My knowledge of theory is unfortunately weak, but somewhere I've read that it is important to correct the df for the Q-statistics when they are computed for the residuals from an ARMA estimation. I really don't know what that means, so kindly help me understand if what I am doing is wrong. Thank you.
Working...
X