Vector generalized additive models.

vgam-class

An implementation of about 6 major classes of
statistical regression models. The central algorithm is
Fisher scoring and iterative reweighted least squares.
At the heart of this package are the vector generalized linear
and additive model (VGLM/VGAM) classes. VGLMs can be loosely
thought of as multivariate GLMs. VGAMs are data-driven
VGLMs that use smoothing. The book "Vector Generalized
Linear and Additive Models: With an Implementation in R"
(Yee, 2015) <DOI:10.1007/978-1-4939-2818-7> gives details of
the statistical framework and the package. Currently only
fixed-effects models are implemented. Many (100+) models and
distributions are estimated by maximum likelihood estimation
(MLE) or penalized MLE. The other classes are RR-VGLMs
(reduced-rank VGLMs), quadratic RR-VGLMs, doubly constrained
RR-VGLMs, quadratic RR-VGLMs, reduced-rank VGAMs,
RCIMs (row-column interaction models)---these classes perform
constrained and unconstrained quadratic ordination (CQO/UQO)
models in ecology, as well as constrained additive ordination
(CAO). Hauck-Donner effect detection is implemented.
Note that these functions are subject to change;
see the NEWS and ChangeLog files for latest changes.

Thomas Yee

VGAM

Vector Generalized Linear and Additive Models

Cleve Moler

vgam-class function

Class ``vgam'' — vgam-class

Objects can be created by calls of the form <code>vgam(...)</code>.

Objects from the Class

<dl>
 <dt><code>nl.chisq</code>:</dt>
<dd>Object of class <code>"numeric"</code>.
Nonlinear chi-squared values.</dd>

 <dt><code>nl.df</code>:</dt>
<dd>Object of class <code>"numeric"</code>.
Nonlinear chi-squared degrees of freedom values.</dd>

 <dt><code>spar</code>:</dt>
<dd>Object of class <code>"numeric"</code>
containing the (scaled) smoothing parameters.</dd>

 <dt><code>s.xargument</code>:</dt>
<dd>Object of
 class <code>"character"</code>
holding the variable name of any <code>s()</code> terms.</dd>

 <dt><code>var</code>:</dt>
<dd>Object of class <code>"matrix"</code> holding
 approximate pointwise standard error information.</dd>

 <dt><code>Bspline</code>:</dt>
<dd>Object of class <code>"list"</code>
 holding the scaled (internal and boundary) knots, and the
 fitted B-spline coefficients. These are used
for prediction.</dd>

 <dt><code>extra</code>:</dt>
<dd>Object of class <code>"list"</code>;
the <code>extra</code> argument on entry to <code>vglm</code>. This
contains any extra information that might be needed
 by the family function.</dd>

 <dt><code>family</code>:</dt>
<dd>Object of class <code>"vglmff"</code>.
The family function.</dd>

 <dt><code>iter</code>:</dt>
<dd>Object of class <code>"numeric"</code>.
The number of IRLS iterations used.</dd>

 <dt><code>predictors</code>:</dt>
<dd>Object of class <code>"matrix"</code>
 with \(M\) columns which holds
 the \(M\) linear predictors.</dd>

 <dt><code>assign</code>:</dt>
<dd>Object of class <code>"list"</code>,
 from class <code> "vlm"</code>.
 This named list gives information matching
 the columns and the
 (LM) model matrix terms.</dd>

 <dt><code>call</code>:</dt>
<dd>Object of class <code>"call"</code>,
 from class
 <code> "vlm"</code>.
 The matched call.</dd>

 <dt><code>coefficients</code>:</dt>
<dd>Object of class
<code>"numeric"</code>, from class <code> "vlm"</code>.
A named vector of coefficients.</dd>

 <dt><code>constraints</code>:</dt>
<dd>Object of
 class <code>"list"</code>, from
 class <code> "vlm"</code>.
 A named list of constraint matrices used in the fitting.</dd>

 <dt><code>contrasts</code>:</dt>
<dd>Object of
 class <code>"list"</code>, from
 class <code> "vlm"</code>.
 The contrasts used (if any).</dd>

 <dt><code>control</code>:</dt>
<dd>Object of class <code>"list"</code>,
 from class
 <code> "vlm"</code>.
 A list of parameters for controlling the fitting process.
 See <code>vglm.control</code> for details.</dd>

 <dt><code>criterion</code>:</dt>
<dd>Object of
 class <code>"list"</code>, from
 class <code> "vlm"</code>.
 List of convergence criterion evaluated at the
 final IRLS iteration.</dd>

 <dt><code>df.residual</code>:</dt>
<dd>Object of class
<code>"numeric"</code>, from class <code> "vlm"</code>.
The residual degrees of freedom.</dd>

 <dt><code>df.total</code>:</dt>
<dd>Object of class <code>"numeric"</code>,
 from class <code> "vlm"</code>.
The total degrees of freedom.</dd>

 <dt><code>dispersion</code>:</dt>
<dd>Object of class <code>"numeric"</code>,
 from class <code> "vlm"</code>.
The scaling parameter.</dd>

 <dt><code>effects</code>:</dt>
<dd>Object of class <code>"numeric"</code>,
 from class <code> "vlm"</code>.
The effects.</dd>

 <dt><code>fitted.values</code>:</dt>
<dd>Object of class
 <code>"matrix"</code>, from class <code> "vlm"</code>.
 The fitted values. This is usually the mean but may be
 quantiles, or the location parameter,
 e.g., in the Cauchy model.
</dd>

 <dt><code>misc</code>:</dt>
<dd>Object of class <code>"list"</code>,
 from class <code> "vlm"</code>.
A named list to hold miscellaneous parameters.</dd>

 <dt><code>model</code>:</dt>
<dd>Object of class <code>"data.frame"</code>,
 from class <code> "vlm"</code>.
The model frame.</dd>

 <dt><code>na.action</code>:</dt>
<dd>Object of class <code>"list"</code>,
 from class <code> "vlm"</code>.
A list holding information about missing values.</dd>

 <dt><code>offset</code>:</dt>
<dd>Object of class <code>"matrix"</code>,
 from class <code> "vlm"</code>.
If non-zero, a \(M\)-column matrix of offsets.</dd>

 <dt><code>post</code>:</dt>
<dd>Object of class <code>"list"</code>,
 from class <code> "vlm"</code>
where post-analysis results may be put.</dd>

 <dt><code>preplot</code>:</dt>
<dd>Object of class <code>"list"</code>,
 from class <code> "vlm"</code>
used by <code>plotvgam</code>; the plotting parameters
 may be put here.</dd>

 <dt><code>prior.weights</code>:</dt>
<dd>Object of class
<code>"matrix"</code>, from class <code> "vlm"</code>
holding the initially supplied weights.</dd>

 <dt><code>qr</code>:</dt>
<dd>Object of class <code>"list"</code>,
 from class <code> "vlm"</code>.
QR decomposition at the final iteration.</dd>

 <dt><code>R</code>:</dt>
<dd>Object of class <code>"matrix"</code>,
 from class <code> "vlm"</code>.
 The R matrix in the QR decomposition used
 in the fitting.</dd>

 <dt><code>rank</code>:</dt>
<dd>Object of class <code>"integer"</code>,
 from class <code> "vlm"</code>.
Numerical rank of the fitted model.</dd>

 <dt><code>residuals</code>:</dt>
<dd>Object of class <code>"matrix"</code>,
 from class <code> "vlm"</code>.
The working residuals at the final IRLS iteration.</dd>

 <dt><code>ResSS</code>:</dt>
<dd>Object of class <code>"numeric"</code>,
 from class <code> "vlm"</code>.
Residual sum of squares at the final IRLS iteration with
the adjusted dependent vectors and weight matrices.


</dd>

 <dt><code>smart.prediction</code>:</dt>
<dd>Object of class
<code>"list"</code>, from class <code> "vlm"</code>.
A list of data-dependent parameters (if any)
 that are used by smart prediction.


</dd>

 <dt><code>terms</code>:</dt>
<dd>Object of class <code>"list"</code>,
 from class <code> "vlm"</code>.
The <code><a href='https://rdrr.io/r/stats/terms.html'>terms</a></code> object used.</dd>

 <dt><code>weights</code>:</dt>
<dd>Object of class <code>"matrix"</code>,
 from class <code> "vlm"</code>.
The weight matrices at the final IRLS iteration.
 This is in matrix-band form.</dd>

 <dt><code>x</code>:</dt>
<dd>Object of class <code>"matrix"</code>,
 from class <code> "vlm"</code>.
The model matrix (LM, not VGLM).</dd>

 <dt><code>xlevels</code>:</dt>
<dd>Object of class <code>"list"</code>,
 from class <code> "vlm"</code>.
The levels of the factors, if any, used in fitting.</dd>

 <dt><code>y</code>:</dt>
<dd>Object of class <code>"matrix"</code>,
 from class <code> "vlm"</code>.
The response, in matrix form.</dd>

 <dt><code>Xm2</code>:</dt>
<dd>Object of class <code>"matrix"</code>,
 from class <code> "vlm"</code>.
 See <code>vglm-class</code>).</dd>

 <dt><code>Ym2</code>:</dt>
<dd>Object of class <code>"matrix"</code>,
 from class <code> "vlm"</code>.
 See <code>vglm-class</code>).</dd>

 <dt><code>callXm2</code>:</dt>
<dd>Object of class <code>"call"</code>, from class <code> "vlm"</code>.
 The matched call for argument <code>form2</code>.</dd>

 
</dl>

Slots

Class <code>"vglm"</code>, directly.
Class <code>"vlm"</code>, by class <code>"vglm"</code>.

Extends

<dl>
 <dt>cdf</dt>
<dd><code>signature(object = "vglm")</code>:
cumulative distribution function.
Useful for quantile regression and extreme value data models.</dd>

 <dt>deplot</dt>
<dd><code>signature(object = "vglm")</code>:
 density plot.
 Useful for quantile regression models.</dd>

 <dt>deviance</dt>
<dd><code>signature(object = "vglm")</code>:
deviance of the model (where applicable).</dd>

 <dt>plot</dt>
<dd><code>signature(x = "vglm")</code>:
diagnostic plots.</dd>

 <dt>predict</dt>
<dd><code>signature(object = "vglm")</code>:
extract the additive predictors or
predict the additive predictors at a new data frame.</dd>

 <dt>print</dt>
<dd><code>signature(x = "vglm")</code>:
short summary of the object.</dd>

 <dt>qtplot</dt>
<dd><code>signature(object = "vglm")</code>:
quantile plot (only applicable to some models).</dd>

 <dt>resid</dt>
<dd><code>signature(object = "vglm")</code>:
residuals. There are various types of these.</dd>

 <dt>residuals</dt>
<dd><code>signature(object = "vglm")</code>:
residuals. Shorthand for <code>resid</code>.</dd>

 <dt>rlplot</dt>
<dd><code>signature(object = "vglm")</code>:
 return level plot.
Useful for extreme value data models.</dd>

 <dt>summary</dt>
<dd><code>signature(object = "vglm")</code>:
a more detailed summary of the object.</dd>

 
</dl>

vgam-class: Class ``vgam''

Description

Arguments

References

See Also

Examples