.loss

Calculate loss from predicted and observed values

Implements stacked elastic net regression (Rauschenberger 2021, <doi:10.1093/bioinformatics/btaa535>). The elastic net generalises ridge and lasso regularisation (Zou 2005, <doi:10.1111/j.1467-9868.2005.00503.x>). Instead of fixing or tuning the mixing parameter alpha, we combine multiple alpha by stacked generalisation (Wolpert 1992 <doi:10.1016/S0893-6080(05)80023-1>).

Armin Rauschenberger

starnet

Stacked Elastic Net

.loss function

<dl><dt>y</dt>
<dd>observed values:
numeric vector of length \(n\)</dd>
<dt>x</dt>
<dd>predicted values:
numeric vector of length \(n\)</dd>
<dt>family</dt>
<dd>character <code>"gaussian"</code>, <code>"binomial"</code>, <code>"poisson"</code>,
<code>"mgaussian"</code>, or <code>"multinomial"</code> (to implement: <code>"cox"</code>)</dd>
<dt>type.measure</dt>
<dd>character <code>"deviance"</code>, <code>"mse"</code>, <code>"mae"</code>, <code>"class"</code>,
or <code>"auc"</code></dd>
<dt>foldid</dt>
<dd>fold identifiers:
integer vector of length \(n\),
or <code>NULL</code></dd>
<dt>grouped</dt>
<dd>logical (for <code>"cox"</code> only)</dd></dl>

Arguments

Loss — .loss

<dl>

<dt>y</dt>
<dd>observed values:
numeric vector of length \(n\)</dd>


<dt>x</dt>
<dd>predicted values:
numeric vector of length \(n\)</dd>


<dt>family</dt>
<dd>character <code>"gaussian"</code>, <code>"binomial"</code>, <code>"poisson"</code>,
<code>"mgaussian"</code>, or <code>"multinomial"</code> (to implement: <code>"cox"</code>)</dd>


<dt>type.measure</dt>
<dd>character <code>"deviance"</code>, <code>"mse"</code>, <code>"mae"</code>, <code>"class"</code>,
or <code>"auc"</code></dd>


<dt>foldid</dt>
<dd>fold identifiers:
integer vector of length \(n\),
or <code>NULL</code></dd>


<dt>grouped</dt>
<dd>logical (for <code>"cox"</code> only)</dd>

</dl>

.loss: Loss

Description

Usage

Arguments

Examples