generalized ordered logit in R or python

user_8838308 0
Viewed 2100

I would like to fit a generalized ordered logit model to some data I have. I first tried to use the ordered logit model using the MASS package from R, but it seems that the proportional odds assumption is violated by the data. Indeed, not all independent variables do exert the same effect across all categories of the dependent variable. Hence I am a bit blocked. I say that I could use the generalized ordered logit model instead, but could not find how to use it. Indeed, I can not find any package on either R or python that coud help me on that.

If someone has any hints on packages I could use, it would be of great help!

Thank you very much!

logistic-regression r python
2 Answers

Try the oglmx package. In this link there is the vignette for that package.

For those who are interested in the method itself, here is some talk about that.

user_88381690

Try the 'VGAM' package. There is a function called vglm. example: vglm(Var2~factor(Var1),cumulative(parallel = F),data) generalized order model cumulative(parallel=T) will perform a proportional odds model. parallel=F is the default.

user_122864370
Related