The RATS Software Forum

Hello.

I'm using RATs to model several stochastic regressors and need a way of pre whitening MA terms. The FILTER instruction only supports AR and distributed lag options for equations. Does anyone know of any user procedures that support MA filtering using RATS?

Thanks.

Scott

sbriggs wrote:Hello.

I'm using RATs to model several stochastic regressors and need a way of pre whitening MA terms. The FILTER instruction only supports AR and distributed lag options for equations. Does anyone know of any user procedures that support MA filtering using RATS?

Thanks.

Scott

What is it that you're trying to do? Are you estimating an MA model for a series and want to apply that same filter to other series?

Yes, essentially.

I have modeled a series that contains both AR and MA terms. Now I need to pre whiten the dependent variable (DV) as I want to estimate a transfer function after cross correlating the two series. I can use the FILTER instruction to pre whiten the DV for the AR term(s) but not for the MA term(s). Looking for a way to do the MA filtering.

Thanks.

Scott

sbriggs wrote:Yes, essentially.

I have modeled a series that contains both AR and MA terms. Now I need to pre whiten the dependent variable (DV) as I want to estimate a transfer function after cross correlating the two series. I can use the FILTER instruction to pre whiten the DV for the AR term(s) but not for the MA term(s). Looking for a way to do the MA filtering.

Thanks.

Scott

You can do that with BOXJENK by feeding in the coefficients from the BOXJENK that you've estimated. See the delup562.rpf example in the Delurgio textbook examples.