Exhaustive Model Selection using One or More Factor Variables in R

Question

I am attempting to construct a Linear Model based on a large number of explanatory variables (including interaction terms between variables) two of which are factor variables. Using a factor variable in either regsubsets() or leaps() causes an error. Is there another function or package I can use that will incorporate factor variables when attempting exhaustive model selection?

Thank You

James

yep,came here to suggest `glmulti` too but Ben beat me to it. — Maiasaura, Sep 15 '12 at 17:43

score 1 · Accepted Answer · edited Nov 08 '17 at 12:52

1

I've had the same issue when setting up the leaps() or regsubsets() using the x and y matrices.

However, regsubsets(y~x1cont+factor(x2)..., data=data) works for me. No error. Perhaps that may work for you?

edited Nov 08 '17 at 12:52

Ryan

26,884
9
56
83

answered Sep 25 '12 at 15:36

Moose

26
1

Exhaustive Model Selection using One or More Factor Variables in R

1 Answers1