how vifcor function works in R

Question

I'm using vifcor and vifstep functions from the usdm package in R to detect multicollinearity. My understanding for vifcor is that if I put the threshold as 0.9 for example it should give me all the variables with vif values <= 9. But the results showed much higher values (39, etc.).

So how exactly does it work and what is its relation to the vif value?

vif

If you look at the documentation for vifcor you can read about the `th` argument and what it does. I don't think there is necessarily any connection between the correlation of two variables in the dataset (max. 1) and the associated VIF (not bounded) — qdread, Oct 27 '20 at 00:52

score 1 · Accepted Answer · answered Oct 27 '20 at 00:58

1

For vifcor, th is filtering based on the correlation threshold not VIF. Use vifstep if you want to filter using VIF.

answered Oct 27 '20 at 00:58

Paul

8,734
1
26
36

how vifcor function works in R

1 Answers1