[R-sig-eco] cozigam
Ben Bolker
bbolker at gmail.com
Wed Jun 13 15:48:30 CEST 2012
On 12-06-13 08:27 AM, Mahnaz Rabbaniha wrote:
> thanks
>
> in continue
>
> your idea exactly correct, i study this lecture and i know for checking
> of model is necessary do diagnostic
> plots including the Q-Q normal score plot of the residuals and the plot
> of residuals vs. tted
> values , but do you introduce which code is usable?
I don't know. You can use the standard tools (qqnorm(), plot(),
etc.), use the predict() method to get predicted values of the mean and
zero-inflated part, and compute residuals from that. I don't know
offhand how to fit the zero-inflation probability and the predicted mean
of the non-zero part together, but I imagine that the latter part (the
Gaussian part of the model) should be well-behaved for the non-zero
responses. (i.e., take the non-zero values of your response variable,
subtract the corresponding predictions to get residuals).
Maybe look at the new Zuur et al. book on zero-inflated models for
guidance?
>
> thanks
>
> On Wed, Jun 13, 2012 at 4:10 PM, Ben Bolker <bbolker at gmail.com
> <mailto:bbolker at gmail.com>> wrote:
>
> [cc'ing back to r-ecology]
>
>
> -------- Original Message --------
> Subject: Re: [R-sig-eco] cozigam
> Date: Wed, 13 Jun 2012 14:38:49 +0430
> From: Mahnaz Rabbaniha <rab.mahnaz at gmail.com
> <mailto:rab.mahnaz at gmail.com>>
> To: Ben Bolker <bbolker at gmail.com <mailto:bbolker at gmail.com>>
>
> hi
>
> thanks for your answer
>
> for finding the relation i have try glm,gam and gam with smooth
> variable, but in all conditions the results shown unacceptable answer (
> for example: R-sq.(adj) = 0.1 , Deviance explained = 18.9%)
>
> in base of contacts previous whit r - group and in base of zero in
> data,i decided to use cozigam,My awareness is low about it but i try
> different code in Liu,2010 .
>
> after the received mis mentioned above, i omitted depth and this
> code used:
>
> res <- cozigam(Clupeidae~s(temperature,salinity), constraint =
> "proportional", family = gaussian)
>
> result:
>
> [snip]
>
> > summary(res)
> Family: gaussian
> Parametric coefficients:
> Estimate Std. Error t value Pr(>|t|)
> (Intercept) -20.665505 13.904231 -1.486 0.1401
> alpha -0.486186 0.207679 -2.341 0.0210 *
> delta1 0.010161 0.004781 2.125 0.0358 *
>
> Approximate significance of smooth terms:
>
> Edf Est.rank F p-value
>
> s(temperature,salinity) 20.81 29 8.77
> <2e-16 ***
> ---
>
> Scale est. = 270.56 n = 132
>
> what do you think? is it adequate for analyses ? do you have any suggest
>
>
> BMB> This is really too vague a question. You should do the usual
> things that are done with the results of any analysis: figure out what
> the parameters mean (e.g. by reading the JRSS COZIGAM paper:
> http://www.jstatsoft.org/v35/i11/paper ), look at the parameter
> estimates, their confidence intervals, predictions and see if they
> make sense, residuals and see if there are obvious violations of
> the statistical models (systematic patterns, variation in heterogeneity,
> etc.)
>
> On Wed, Jun 13, 2012 at 1:03 PM, Ben Bolker <bbolker at gmail.com
> <mailto:bbolker at gmail.com>
> <mailto:bbolker at gmail.com <mailto:bbolker at gmail.com>>> wrote:
>
> Mahnaz Rabbaniha <rab.mahnaz at ...> writes:
>
> > i try to find regression between clupeidae,with
> temperature,salinity and
> > depth. the response variable is inclued many zero ( 86 from 133
> observed)
> >
> > therefore i used this code :
> >
> > res <- cozigam(Clupeidae~s(temperature,salinity)+s(depth),
> constraint =
> > "proportional", family = gaussian)
> >
> > the result:
> > iteration = 2 norm = 1.001743
> > iteration = 3 norm = 0.3377464
> > iteration = 4 norm = 9.172232e-05
> >
> > ==========================================
> > estimated alpha = -0.5337883 ( 0.1789113 )
> > estimated delta = -0.0009891505 ( NaN )
> > ==========================================
> >
> > Warning message:
> >
> > In sqrt(V.theta[2, 2]) : NaNs produced
> >
> > what is exactly meaning?
>
>
> You're probably not getting answers to your repeated posts
> because you're not providing a reproducible example
> ( http://tinyurl.com/reproducible-000 ) and not giving very much
> detail about your problem.
> I strongly suspect that your model is too complex for your data:
> a general rule of thumb is that you need about 10 observations
> per parameter estimated. It's a bit hard to count in this case
> for two reasons -- zeroes are relatively uninformative (so each
> zero counts for less than one 'effective' observation), and it's
> a little hard to count parameters for penalized smooth terms --
> but I think you can't really expect to fit a two-way smooth term
> on temperature and salinity *and* a smooth term on depth ... the
> example in the COZIGAM JRSS paper (referenced in the help)
> fits a model of about the same complexity to 274 data points with
> 84 zero catches -- somewhere between 3 and 4 times as much data
> as you have.
> Most narrowly, the program is trying to estimate the standard
> error of the parameter by inverting the matrix of second derivatives,
> and failing because the surface is too flat, or too strongly
> correlated, or some similar problem.
>
> _______________________________________________
> R-sig-ecology mailing list
> R-sig-ecology at r-project.org <mailto:R-sig-ecology at r-project.org>
> <mailto:R-sig-ecology at r-project.org
> <mailto:R-sig-ecology at r-project.org>>
> https://stat.ethz.ch/mailman/listinfo/r-sig-ecology
>
>
>
>
>
>
>
>
>
>
>
> --
> Mahnaz Rabbaniha
> Senior expert of marine ecology
> Iranian Fisheries Research Organization (IFRO)
> P.O.Box: 14155-6116 , P.Code: 1411816618*
> Tehran, IRAN
> Phone: +98 21 44580953
> Fax: +98 21 44580583
> Mobile: +98 912 5790377
>
>
>
>
>
r
More information about the R-sig-ecology
mailing list