0
$\begingroup$

Data set: response and predictors are all non-stationary, time series variables

After performing Box-Cox transformations and testing a variety of power transformations on each variable, the non-stationarity still exists. Linear models have therefore been discarded in favor of autoregressive models.

To construct ARIMA models, the variables were differenced, but the acf and pacf plots still have bars that exceed the significance bounds at all difference levels.

What statistical approach is next in line? Nonlinear transformations? Nonlinear models?

$\endgroup$
  • $\begingroup$ Could you provide a sample of your data and/or plots so to illustrate your question? This would help us to get better picture of your problem. $\endgroup$ – Tim Apr 14 '15 at 13:54
  • $\begingroup$ If the ACF and PACF bars did not exceed the significance bounds, there would be no use for the ARIMA model. Having significant bars is thus desirable if you want to proceed with an ARIMA model. $\endgroup$ – Richard Hardy Apr 14 '15 at 13:54

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.