Skip to content
Prev 298393 / 398498 Next

Zero inflated: is there a limit to the level of inflation

Thank you both for your quick response and input. I will consider all of
your points and see what we are able to derive from there. 

Thank you again for your time and expertise.

-Stephanie

-------------------------------------------------------
Stephanie L. Simek
Carnivore Ecology Lab
Forest and Wildlife Research Center
Mississippi State University
Box 9690
Mississippi State, MS 39762
Cell: (850) 591-1430
Email: ssimek at cfr.msstate.edu


-----Original Message-----
From: Achim Zeileis [mailto:Achim.Zeileis at uibk.ac.at] 
Sent: Tuesday, June 26, 2012 4:46 PM
To: Marc Schwartz
Cc: Stephanie L. Simek; r-help at r-project.org
Subject: Re: [R] Zero inflated: is there a limit to the level of
inflation
On Tue, 26 Jun 2012, Marc Schwartz wrote:

            
individual:
Season	Cov
AUT	Open
SPR	Open
SUM	Open
SUM	Open
AUT	oldHard
SPR	oldHard
SUM	oldHard
SUM	oldHard
AUT	Water
SPR	Water
SUM	Water
SUM	Open
SUM	oldHard
medHard
SUM	Water
SUM	Water
AUT	Open
AUT	Open
SPR	Open
SPR	Open
SUM	Open
AUT	oldHard
AUT	oldHard
SPR	oldHard
SPR	oldHard
2005	AUT	Water
2006	AUT	Water
2006	SPR	Water
2007	SPR	Water
2006	SUM	Water
was monitored.
negative binomial distribution rather than zero inflated poisson.
which provides standard fixed effects models and related functions for
count based data and importantly, some good conceptual content:
likely have, you should subscribe to and re-post your query to the
R-sig-mixed-models list:
in the above list. Based upon the content there, I suspect that you will
be pointed to the glmmADMB package which is on R-Forge
(http://glmmadmb.r-forge.r-project.org/) and can handle zero inflated
mixed effects models of at least some types.
data.
Thanks, Marc, all very useful points! Just one addition:

I would recommend starting with the last point - a binary response
regression (for y > 0). This could be considered as the zero-hurdle of a
hurdle regression.

Hurdle regressions are an alternative to zero-inflated models, but have
the nice property that you can separately estimate both parts of the
hurdle: (1) a binary regression for y=0 vs. y > 0. (2) A truncated count
model for y, estimated only from the observations y>0. The "pscl"
package contains a hurdle() function which estimates both parts in one
go (and the "countreg" vignette gives more details and references), but
in this case it would probably be useful to estimate them separately.

In any case, both parts will need care because the binary response
probably contains a lot of (quasi-)complete separations because
non-zeros are so rare. Conversely, the truncated count model may be hard
to estimate because there are no observations for a lot of parameter
combinations. But estimating the models separately will give you more
flexibility in addressing these issues.

To estimate the zero-truncated count distributions, you may consider the
"countreg" package from R-Forge which uses the same code as (one part
of) the hurdle() function.

hth,
Z