[R-sig-ME] Zero-inflated and proportional data

[Marc Schwartz]

> On Nov 14, 2014, at 5:41 PM, Mayara Amescua <may.amescua at gmail.com> wrote:
> 
> I need to apply a zero-inflated regression to a proportional (percentage)
> data and I can't find a way to do that. Is there anyone who knows if it's
> possible?


Hi,

What you seem to essentially want is "zero inflated beta regression", since standard beta regression operates on the unit interval (0, 1) and thus, cannot handle 0's and 1's. 

A Google search using the above term brings up various references.

I have not used them, but take a look at the GAMLSS package:

  http://cran.r-project.org/web/packages/gamlss/

and the zoib package:

  http://cran.r-project.org/web/packages/zoib/

Both have reference materials available via the CRAN pages and/or the author's web sites.

Also, it is not clear from your query that this is the correct list, which concerns mixed (random and fixed effects) models. If not, post any follow ups to the regular R-Help list or perhaps one of the other SIG lists as may be apropos for the subject matter domain:

  http://www.r-project.org/mail.html

Regards,

Marc Schwartz