bachlaw01 at outlook.com
Sun Feb 25 23:55:51 CET 2018
The pscl package offers the (somewhat controversial) Vuong test for this purpose and is a good ZI/hurdle resource in general.
> On Feb 25, 2018, at 3:45 PM, C. AMAL D. GLELE <altessedac2 at gmail.com> wrote:
> Hi, Ben
> Many thanks to you for your very helpful reply.
> Best and regards,
> 2018-02-25 19:05 GMT+01:00 Ben Bolker <bbolker at gmail.com>:
>> There is no set proportion. (For example, a Poisson distribution with
>> a mean of 0.01 is expected to be about 99% zeros, even without
>> zero-inflation.) There's a little bit of (bare-bones) discussion of
>> how to test for zero-inflation here:
>> On Sun, Feb 25, 2018 at 7:08 AM, C. AMAL D. GLELE <altessedac2 at gmail.com>
>>> Hi, dear all.
>>> From which proportion of zero a count should be considered as
>>> (in order to use a zero-inflated model for it's modelling)?
>>> In advance, thanks for your replies.
>>> sans virus. www.avast.com
>>> [[alternative HTML version deleted]]
>>> R-sig-mixed-models at r-project.org mailing list
> [[alternative HTML version deleted]]
> R-sig-mixed-models at r-project.org mailing list
More information about the R-sig-mixed-models