I suppose there may be an error of sorts, but have you considered
the fact that solving the error might not gain you admittance into
heaven? Look at the RHS of the model:
sensor2 + s(site, bs = re)
... and think about the fact that you are smoothing a factor
variable.
- Actually this is ok.
Hi Katharina,
Thanks for sending this.
The problem is that the prediction data for site contain levels not
available in the (useable non-NA) fit data...
levels(m$model$site)
[1] KRB NP.FOR WKS.FRE WKS.KRE WKS.RIE WKS.WUE
levels(gapData$site)
[1] KRB NP.FOR RIE.2 WKS.BBR WKS.FRE
Hi Simon,
many thanks for looking into this and making me understand the problem!
I'll adjust my factor levels right away...
Best, Katharina
On 3 February 2014 12:42, Simon Wood s.w...@bath.ac.uk wrote:
Hi Katharina,
Thanks for sending this.
The problem is that the prediction data for site
Dear Simon, your note below says bs=re specifies a Gaussian random
effect . I have been using bs = re for data modeled with Poisson and
binomial distributions, or variants thereof (e.g., quasi-Poisson). Have
I erred in assuming bs =re can be used to obtain random effects for
such data? Will
The two distributions are different. The random effect is assumed to
be a Gaussian random variable, just as it is with the GLMMs in the
lme4 package. It is fine to use such a random effect within a GAM with
a non-Gaussian error distribution, like the ones you describe using.
HTH
Gavin
On 3
Hi Simon,
thank you for your reply, I really appreciate any help to understand
the problem here...
Unluckily the package upgrade didn't help with this issue.
An example reproducing the error, and a current sessionInfo() Output
can be found below.
Many thanks once again,
Katharina
R
On Feb 2, 2014, at 9:52 AM, Katharina May wrote:
Hi Simon,
thank you for your reply, I really appreciate any help to understand
the problem here...
Unluckily the package upgrade didn't help with this issue.
An example reproducing the error, and a current sessionInfo() Output
can be found
Hi Katharina,
Could you try upgrading to mgcv_1.7-28, please? There was an occasional
problem to do with matching factor levels, which is fixed, but I'm not
very confident that is what is going on.
If upgrading doesn't work, is there any chance you could send me a small
example dataset and
Dear R-Community,
I`m trying to apply the mgcv package to fill gaps in sensor data from
different sites (9 sites, 2 sensors per site) and do the filling on a
site-wise level.
Based on
http://r.789695.n4.nabble.com/mgcv-gamm-predict-to-reflect-random-s-effects-td3622738.html
my model looks like
9 matches
Mail list logo