ou again. I appreciate your input.
>
> -Nick
> --------------------
> On Fri, 8/29/14, Achim Zeileis wrote:
>
> Subject: Re: [R] Question regarding the discrepancy between count model
> parameter estimates between "pscl" and "MASS"
>
st did so in the GUI: just untick 'binary'.)
Thank you again. I appreciate your input.
-Nick
On Fri, 8/29/14, Achim Zeileis wrote:
Subject: Re: [R] Question regarding the discrepancy between count model parameter estimates
between
ct: Re: [R] Question regarding the discrepancy between count model
parameter estimates between "pscl" and "MASS"
To: "peter dalgaard"
Cc: "Nick Livingston" , r-help@r-project.org
Date: Friday, August 29, 2014, 5:26 AM
On Fri, 29 Aug 2014,
peter dalgaard wrote
On Fri, 29 Aug 2014, peter dalgaard wrote:
I'm no expert on hurdle models, but it seems that you are unaware that
the negative binomial and the truncated negative binomial are quite
different things.
Yes. You can replicate the truncated count part of the hurdle model with
the zerotrunc() fun
I'm no expert on hurdle models, but it seems that you are unaware that the
negative binomial and the truncated negative binomial are quite different
things.
-pd
On 29 Aug 2014, at 05:57 , Nick Livingston wrote:
> I have sought consultation online and in person, to no avail. I hope someone
>
I have sought consultation online and in person, to no avail. I hope someone
on here might have some insight. Any feedback would be most welcome.
I am attempting to plot predicted values from a two-component hurdle model
(logistic [suicide attempt yes/no] and negative binomial count [number of
att
6 matches
Mail list logo
