bootstrap_rmse

<code>bootstrap_rmse()</code> is deprecated in favour of <code>bootstrap_performance()</code>.
Objects returned by <code>bootstrap_rmse()</code> keep class <code>"bootstrap_rmse"</code> for
backward compatibility and also inherit from <code>"bootstrap_performance"</code>.

internal

Provides actuarial tools and building blocks for analysing,
modelling, refining, and validating insurance rating models. Designed to
support common GLM-based pricing tasks and the translation of statistical
model output into practical tariff structures. The package supports the
construction of insurance tariff classes using a data-driven approach, based
on the methodology of Antonio and Valdez
(2012) <doi:10.1007/s10182-011-0152-7>.

Martin Haringa

insurancerating

Actuarial Tools for Insurance Pricing Models

bootstrap_rmse function

<dl><dt>model</dt>
<dd>A fitted model object that can be updated on resampled data.</dd>
<dt>data</dt>
<dd>Data frame containing the model response and predictors.</dd>
<dt>n</dt>
<dd>Deprecated. Use <code>n_resamples</code> in <code>bootstrap_performance()</code> instead.</dd>
<dt>frac</dt>
<dd>Deprecated. Use <code>sample_fraction</code> in <code>bootstrap_performance()</code>
instead.</dd>
<dt>metric</dt>
<dd>Character string. Performance metric to compute. Currently only
<code>"rmse"</code> is supported.</dd>
<dt>sampling</dt>
<dd>Character string. Sampling scheme. <code>"bootstrap"</code> samples
training
rows with replacement and evaluates on out-of-bag rows when
<code>sample_fraction &lt; 1</code>.
<code>"split"</code> samples training rows without replacement and evaluates on the
remaining rows when <code>sample_fraction &lt; 1</code>.</dd>
<dt>show_progress</dt>
<dd>Logical. Show a progress bar during resampling.
Default is <code>TRUE</code>.</dd>
<dt>rmse_model</dt>
<dd>Optional finite numeric RMSE for the original fitted model.
If <code>NULL</code>, it is calculated from <code>model</code> and <code>data</code>.</dd></dl>

Arguments

Deprecated alias for bootstrap_performance() — bootstrap_rmse

<dl>

<dt>model</dt>
<dd>A fitted model object that can be updated on resampled data.</dd>


<dt>data</dt>
<dd>Data frame containing the model response and predictors.</dd>


<dt>n</dt>
<dd>Deprecated. Use <code>n_resamples</code> in <code>bootstrap_performance()</code> instead.</dd>


<dt>frac</dt>
<dd>Deprecated. Use <code>sample_fraction</code> in <code>bootstrap_performance()</code>
instead.</dd>


<dt>metric</dt>
<dd>Character string. Performance metric to compute. Currently only
<code>"rmse"</code> is supported.</dd>


<dt>sampling</dt>
<dd>Character string. Sampling scheme. <code>"bootstrap"</code> samples
training
rows with replacement and evaluates on out-of-bag rows when
<code>sample_fraction &lt; 1</code>.
<code>"split"</code> samples training rows without replacement and evaluates on the
remaining rows when <code>sample_fraction &lt; 1</code>.</dd>


<dt>show_progress</dt>
<dd>Logical. Show a progress bar during resampling.
Default is <code>TRUE</code>.</dd>


<dt>rmse_model</dt>
<dd>Optional finite numeric RMSE for the original fitted model.
If <code>NULL</code>, it is calculated from <code>model</code> and <code>data</code>.</dd>

</dl>

bootstrap_rmse: Deprecated alias for `bootstrap_performance()`

Description

Usage

Value

Arguments