The variable is log return, which has been confirmed as a stationary series by the DF test, I choose the case of p and q are 1 is the best ARMA model, and then at the bottom after entering predict error, resid command, and then ac error, why the graph of ac error and ac log return is exactly the same graph? I later change any p and q but still same, it seems Stata not loaded into my model
Array
Array
0 Response to Problems with the ARIMA model for white noise testing
Post a Comment