02-23-2016 12:26 PM
I need to fit a logistic regression model and am wondering which model-seletion method would be best. I have been advised to stay away from forward/backward/stepwise regression. All-possible-regression seems attractive, but I must admit I'm a little lost on AIC/BIC/Cp/etc and exactly how I would go about picking the best model...
I have a binary response variable, a categorical predictor, 10 categorical covariates, and 2 continuous covariates.
Thank you in advance!
02-23-2016 12:27 PM
Search Model Selection Method on here...this topic comes up frequently, and there is no 'CORRECT' answer, but some answers are more valid than others
02-23-2016 09:12 PM
Unfortunately I've been all over the boards and haven't found anything useful. I've also read several papers - I just can't seem to locate the syntax for an all-possible. In addition, I was hoping someone could break it down for me in less technical language so I could really understand AIC/Cp/etc...
02-24-2016 01:04 PM
After finding the potential factor/variable for inclusion in the model using any of:
- selection = stepwise slentry = 0.15 slstay = 0.15;
- selection = forward slentry =0.15
- selection = backward slstay = 0.15
- selection = score ,
for both quantitative and categorical variables and interaction term - you can compare models based on following criteria:
02-23-2016 03:15 PM
02-23-2016 10:11 PM
You gotta know forward/backward/stepwise regression all these are doing unconditional logistic regression.
After getting the most influent variables , to get Best Fit , you'd better try Exact logistic regression or Conditional logistic regression or Penalty logistic regression(add FIRTH option into ( MODEL statement ) .