topic Re: Autoreg total vs regress R square in Statistical Procedures

Autoreg total vs regress R square

Taliah — Sun, 22 Dec 2019 08:24:33 GMT

In the output of proc autoreg, what is the difference between Regress R square and Total R square?

Which one should be used as an assesment of the goodness of fit of the model?

Re: Autoreg total vs regress R square

JosvanderVelden — Thu, 26 Dec 2019 13:50:34 GMT

Goodness of fit measures and information criteria are described here: https://go.documentation.sas.com/?cdcId=pgmsascdc&cdcVersion=9.4_3.5&docsetId=etsug&docsetTarget=etsug_autoreg_details21.htm&locale=en

Re: Autoreg total vs regress R square

Taliah — Sun, 29 Dec 2019 05:31:08 GMT

Thank you.

After reading the link you gave I still can't fingure out which is the measure to use. From the link you sent:

"The Total Rsq is a measure of how well the next value can be predicted using the structural part of the model and the past values of the residuals. "

" The transformed regression R-square statistic is a measure of the fit of the structural part of the model after transforming for the autocorrelation and is the R-square for the transformed regression."

In the specific autoregression we are using we got Total R sq=0.8, Regress R sq = 0.049. We need to know which to use for estimating the goodness of fit of our regression. I am aware of the other measures of goodness of fit; for other projects we use MAPE, but for this project we need to use and report one of the R squares, and after reading the link, I still can't tell which to use..