topic Re: proc autoreg and a logistic regression equation in Statistical Procedures

proc autoreg and a logistic regression equation

MaryA_Marion — Sun, 23 Apr 2017 20:08:07 GMT

I have been asked to use proc autoreg to compute the durbin watson statistic for a logistic regression problem. How to do?

Thank you. Mary

proc autoreg data=silica;
model y = exp(-(A B C)) / nlag=?;

Re: proc autoreg and a logistic regression equation

PGStats — Mon, 24 Apr 2017 02:39:54 GMT

proc autoreg doesn't do logistic regression. What is the model that you wish to fit?

Re: proc autoreg and a logistic regression equation

PeterClemmensen — Mon, 24 Apr 2017 05:04:58 GMT

Use Proc GENMOD or Proc LOGISTIC instead 🙂

Re: proc autoreg and a logistic regression equation

StatDave — Mon, 24 Apr 2017 15:17:47 GMT

There is no Durbin-Watson test for binomial data. However, you could consider fitting a GEE model in PROC GENMOD and using TYPE=AR or MDEP(1) to examine the estimated correlation between adjacent observations.

Re: proc autoreg and a logistic regression equation

MaryA_Marion — Thu, 04 May 2017 14:39:15 GMT

thank you for your reply. will the autocorrelation between observations be the same as that between residuals?

mary