07-07-2012 10:09 PM
I built one logistic model on my data (my dependent variable includes four nominal category data). I'm wondering if proc logistic provides one way to calculate overall hit rate?
Anyone has experience about this?
Any help will be appreciated...........Thanks
07-09-2012 05:10 PM
You have 4 levels of your dependent variable so you have a nominal logit model, is that what you were after?
For simple logistic regression with binary dependent variable then SAS would give you percent concordant/discordant which is what you're after I think.
You could output the predicted values and then calculate it manually:
output out =pred predprob=(individual);
I don't know anything about your data, but if its a classification problem then proc discrim may be what you're after.