beta regression and AIC (model selection)

Question

I have a problem with Beta regression in R. I need to choose a model using a process similar to stepAIC . I am using stepGAIC from the gamlss package. I get this error message afetr running this code:

model <- betareg(proportion_of_completed_surveys ~ questions + readibility_lexile_survey_questions_expl, data = conference_surveys)

stepGAIC.CH(model,direction="backward", additive=TRUE)

Error in match.arg(model) : 'arg' should be one of “mean”, “precision”

Any solution would be much appreciated.

Peter

Answer 1

I had the same problem, so I wrote a function that could help you.

Try this:

install_github("SergioGarofalo/StepBeta") 
library(StepBeta)
modelStepAIC <- step.beta(model, k = 2)
summary(modelStepAIC)

I hope this could be useful,

Sergio