predict_glmmLasso

<code>glmmLasso()</code> predicts on newdata objects from a glmmLasso object.

A tool that improves the prediction performance of multilevel
regression with post-stratification (MrP) by combining a number of machine
learning methods. For information on the method, please refer to Broniecki,
Wüest, Leemann (2020) ''Improving Multilevel Regression with
Post-Stratification Through Machine Learning (autoMrP)'' in the
'Journal of Politics'. Final pre-print version:
<https://lucasleemann.files.wordpress.com/2020/07/automrp-r2pa.pdf>.

Philipp Broniecki

autoMrP

Improving MrP with Ensemble Learning

Reto Wüest

Lucas Leemann

Florian Schaffner

Hadley Wickham

predict_glmmLasso function

<dl><dt>census</dt>
<dd>Census data. A <code>data.frame</code> whose column names include
<code>L1.x</code>, <code>L2.x</code>, <code>L2.unit</code>, if specified, <code>L2.reg</code> and
<code>pcs</code>, and either <code>bin.proportion</code> or <code>bin.size</code>.</dd>
<dt>m</dt>
<dd>A <code>glmmLasso()</code> object.</dd>
<dt>L1.x</dt>
<dd>Individual-level covariates. A character vector containing the
column names of the individual-level variables in <code>survey</code> and
<code>census</code> used to predict outcome <code>y</code>. Note that geographic unit
is specified in argument <code>L2.unit</code>.</dd>
<dt>lasso.L2.x</dt>
<dd>Lasso context-level covariates. A character vector
containing the column names of the context-level variables in
<code>survey</code> and <code>census</code> to be used by the lasso classifier. If
<code>NULL</code> and <code>lasso</code> is set to <code>TRUE</code>, then lasso uses the
variables specified in <code>L2.x</code>. Default is <code>NULL</code>.</dd>
<dt>L2.unit</dt>
<dd>Geographic unit. A character scalar containing the column
name of the geographic unit in <code>survey</code> and <code>census</code> at which
outcomes should be aggregated.</dd>
<dt>L2.reg</dt>
<dd>Geographic region. A character scalar containing the column
name of the geographic region in <code>survey</code> and <code>census</code> by which
geographic units are grouped (<code>L2.unit</code> must be nested within
<code>L2.reg</code>). Default is <code>NULL</code>.</dd></dl>

Arguments

Predicts on newdata from glmmLasso objects — predict_glmmLasso

<dl>

<dt>census</dt>
<dd>Census data. A <code>data.frame</code> whose column names include
<code>L1.x</code>, <code>L2.x</code>, <code>L2.unit</code>, if specified, <code>L2.reg</code> and
<code>pcs</code>, and either <code>bin.proportion</code> or <code>bin.size</code>.</dd>


<dt>m</dt>
<dd>A <code>glmmLasso()</code> object.</dd>


<dt>L1.x</dt>
<dd>Individual-level covariates. A character vector containing the
column names of the individual-level variables in <code>survey</code> and
<code>census</code> used to predict outcome <code>y</code>. Note that geographic unit
is specified in argument <code>L2.unit</code>.</dd>


<dt>lasso.L2.x</dt>
<dd>Lasso context-level covariates. A character vector
containing the column names of the context-level variables in
<code>survey</code> and <code>census</code> to be used by the lasso classifier. If
<code>NULL</code> and <code>lasso</code> is set to <code>TRUE</code>, then lasso uses the
variables specified in <code>L2.x</code>. Default is <code>NULL</code>.</dd>


<dt>L2.unit</dt>
<dd>Geographic unit. A character scalar containing the column
name of the geographic unit in <code>survey</code> and <code>census</code> at which
outcomes should be aggregated.</dd>


<dt>L2.reg</dt>
<dd>Geographic region. A character scalar containing the column
name of the geographic region in <code>survey</code> and <code>census</code> by which
geographic units are grouped (<code>L2.unit</code> must be nested within
<code>L2.reg</code>). Default is <code>NULL</code>.</dd>

</dl>

predict_glmmLasso: Predicts on newdata from glmmLasso objects

Description

Usage

Value

Arguments