control_rwl

<code>control_rwl</code> sets the default control arguments
for residual learning , <code>type = "rwl"</code>.
The arguments are passed directly to <code>DynTxRegime::rwl()</code> if not
specified otherwise.

Package for learning and evaluating (subgroup) policies via doubly robust loss functions. Policy learning methods include doubly robust blip/conditional average treatment effect learning and sequential policy tree learning. Methods for (subgroup) policy evaluation include doubly robust cross-fitting and online estimation/sequential validation. See Nordland and Holst (2022) <doi:10.48550/arXiv.2212.02335> for documentation and references.

Andreas Nordland

polle

Policy Learning

Klaus Holst

control_rwl function

<dl><dt>moPropen</dt>
<dd>Propensity model of class "ModelObj", see modelObj::modelObj.</dd>
<dt>moMain</dt>
<dd>Main effects outcome model of class "ModelObj".</dd>
<dt>regime</dt>
<dd>An object of class <a href="/link/formula?package=polle&version=1.6.2" data-mini-rdoc="polle::formula">formula</a> specifying the design of the policy/regime.</dd>
<dt>fSet</dt>
<dd>A function or NULL defining subset structure.</dd>
<dt>lambdas</dt>
<dd>Numeric or numeric vector. Penalty parameter.</dd>
<dt>cvFolds</dt>
<dd>Integer. Number of folds for cross-validation of the parameters.
<code>"logit"</code>, <code>"exp"</code>, <code>"hinge"</code>, <code>"sqhinge"</code>, <code>"huber"</code>.</dd>
<dt>kernel</dt>
<dd>The options are <code>"linear"</code>, <code>"poly"</code>, <code>"radial"</code>.</dd>
<dt>kparam</dt>
<dd>Numeric. Kernel parameter</dd>
<dt>responseType</dt>
<dd>Character string. Options are <code>"continuous"</code>,
<code>"binary"</code>, <code>"count"</code>.</dd>
<dt>verbose</dt>
<dd>Integer.</dd></dl>

Arguments

Control arguments for Residual Weighted Learning — control_rwl

<dl>

<dt>moPropen</dt>
<dd>Propensity model of class "ModelObj", see modelObj::modelObj.</dd>


<dt>moMain</dt>
<dd>Main effects outcome model of class "ModelObj".</dd>


<dt>regime</dt>
<dd>An object of class <a href='https://rdrr.io/r/stats/formula.html'>formula</a> specifying the design of the policy/regime.</dd>


<dt>fSet</dt>
<dd>A function or NULL defining subset structure.</dd>


<dt>lambdas</dt>
<dd>Numeric or numeric vector. Penalty parameter.</dd>


<dt>cvFolds</dt>
<dd>Integer. Number of folds for cross-validation of the parameters.
<code>"logit"</code>, <code>"exp"</code>, <code>"hinge"</code>, <code>"sqhinge"</code>, <code>"huber"</code>.</dd>


<dt>kernel</dt>
<dd>The options are <code>"linear"</code>, <code>"poly"</code>, <code>"radial"</code>.</dd>


<dt>kparam</dt>
<dd>Numeric. Kernel parameter</dd>


<dt>responseType</dt>
<dd>Character string. Options are <code>"continuous"</code>,
<code>"binary"</code>, <code>"count"</code>.</dd>


<dt>verbose</dt>
<dd>Integer.</dd>

</dl>

control_rwl: Control arguments for Residual Weighted Learning

Description

Usage

Value

Arguments