How to get prediction of classes in gam?
I have a problem with getting predicted classes from a gam-model using
the gam function in the mgcv-package.
I have a dataset where I have classified the response variable into 5
classes (i. e. 1, 2, 3, 4, 5) and I want to use a gam-model generated on
a
Gavin,
Welcome to the prairies...
I think this thread is getting off track... a few comments on the
autocorrelation model, then bringing it back to the question I was trying to
ask:
I agree completely that we *must* model the time-series because our readings
are not independent and hence we
Jonathan: I wonder if your solution might be as simple as using one of
your model forms but allowing different intercepts for different plots by
modeling them as fixed effects (using a categorical variable for plot).
This would allow your time series model (whatever specification you use)
to
On Thu, 2013-07-18 at 16:13 +0200, Per Bergström wrote:
How to get prediction of classes in gam?
I have a problem with getting predicted classes from a gam-model using
the gam function in the mgcv-package.
I have a dataset where I have classified the response variable into 5
classes (i. e.
On Thu, 2013-07-18 at 14:34 +, Henkelman, Jonathan wrote:
Gavin,
Welcome to the prairies...
Cheers
snip /
I was up examining Zuur last night and found a passing reference to
this problem on page 160 It may be an option to extend the model with
the AR1 correlation, with a random
Hi,
I am a beginner at R and I would like to get information from a matrix,
where the numbers are larger than or equal to 50m . This means that plants
need to be apart 50m. I need to get the selection as a matrix or data
frame. Attached it is an example...
Thanks
--
Kátia Emídio da Silva DSc
