Reputation: 11
I have been using the package glmulti to run an all subsets model selection procedure. I have a large data set and my global model contains 7 candidate predictors. The issue is that I would like to consider just a single interaction between two of my main predictors. When I include an interaction term in the global model function, glmulti ignores it when I use 'level = 1' designating only main effects considered. Does anyone have a way that I can get the procedure to consider main effects and a single interaction?
Here is a small sample:
GlobalModel
rttdfit<- glm(ROCK~AREA+DRATIO+ELEV+SECC+BLDGD+MAXD*CLIM+I(CLIM^2),data=rttd,family=binomial(link="logit"))
Allsubsets selection
rttdASMS<-glmulti(rttdfit, confsetsize = 5, crit = "aic", level = 1, family = binomial("logit"))
weightable(rttdASMS)
Upvotes: 1
Views: 637
Reputation: 11
I was searching around regarding this question and found a link here that may help future readers:
https://vcalcagnoresearch.wordpress.com/package-glmulti/
Upvotes: 1