I forecast my time series price indies data using Automatic ARIMA model in Eviews. This passed unit root testing by proving no unit toots and selected best fitting model as ARMA (4,4) considering first order of data. But when applying residual diagnostics with Q statistic test it rejects the null hypothesis which means there is correlation for residuals. How can I overcome this situation with minimum effort. Refer attached results of the Q test for 15 lags.

More P. M. C. Thakshila's questions See All
Similar questions and discussions