Estimation of GAM smoothing parameters is most stable if
optimization of the smoothness selection score (GCV, GACV, UBRE/AIC, REML, ML etc) 
is outer to the penalized iteratively
re-weighted least squares scheme used to estimate the model given smoothing 
parameters.
This routine optimizes a smoothness selection score in this way. Basically the 
score is evaluated for each trial set of smoothing parameters by
estimating the GAM for those smoothing parameters. The score is minimized
w.r.t. the parameters numerically, using <code>newton</code> (default), <code>bfgs</code>, <code>optim</code> or <code>nlm</code>. Exact
(first and second) derivatives of the score can be used by fitting with
<code>gam.fit3</code>. This
improves efficiency and reliability relative to relying on finite
difference derivatives.
Not normally called directly, but rather a service routine for <code>gam</code>.

gam.outer

Generalized additive (mixed) models, some of their extensions and
other generalized ridge regression with multiple smoothing
parameter estimation by (Restricted) Marginal Likelihood,
Cross Validation and similar, or using iterated nested Laplace
approximation for fully Bayesian inference. See Wood (2025)
<doi:10.1146/annurev-statistics-112723-034249> for an overview.
Includes a gam() function, a wide variety of smoothers, 'JAGS'
support and distributions beyond the exponential family.

Simon Wood

mgcv

Mixed GAM Computation Vehicle with Automatic Smoothness
Estimation

gam.outer function

Minimize GCV or UBRE score of a GAM using `outer' iteration — gam.outer

<dl>
 
<dt>lsp</dt>
<dd>The log smoothing parameters.</dd>


<dt>fscale</dt>
<dd>Typical scale of the GCV or UBRE/AIC score.</dd>


<dt>family</dt>
<dd>the model family.</dd>


<dt>control</dt>
<dd>control argument to pass to <code>gam.fit3</code> etc.</dd>


<dt>method</dt>
<dd>method argument to <code>gam</code> defining the smoothness criterion to use (but depending on whether or not scale known).</dd>


<dt>optimizer</dt>
<dd>The argument to <code>gam</code> defining the numerical optimization method to use.</dd>


<dt>criterion</dt>
<dd>Which smoothness selction criterion to use. One of <code>"UBRE"</code>,
<code>"GCV"</code>, <code>"GACV"</code>, <code>"REML"</code> or <code>"P-REML"</code>.</dd>


<dt>scale</dt>
<dd>Supplied scale parameter. Positive indicates known.</dd>


<dt>gamma</dt>
<dd>The degree of freedom inflation factor for the GCV/UBRE/AIC score.</dd>


<dt>G</dt>
<dd>List produced by <code>mgcv:::gam.setup</code>, containing most of what's
needed to actually fit a GAM.</dd>

<dt>start</dt>
<dd>starting parameter values.</dd>

<dt>nei</dt>
<dd>List specifying neighbourhood structure if NCV used. See <code>gam</code>.</dd>

<dt>...</dt>
<dd>other arguments, typically for passing on to <code>gam.fit3</code> (ultimately).</dd>

</dl>

Arguments

Simon N. Wood <a href='mailto:simon.wood@r-project.org'>simon.wood@r-project.org</a>

gam.outer: Minimize GCV or UBRE score of a GAM using `outer' iteration

Description

Usage

Arguments

Details

References

See Also