compute_qoi_user

Compute Quantities of Interest (PAPE, PAPEp, PAPDp, AUPEC, GATE, GATEcv) with user defined functions

Provides various statistical methods for evaluating
Individualized Treatment Rules under randomized data. The provided
metrics include Population Average Value (PAV), Population Average
Prescription Effect (PAPE), Area Under Prescription Effect Curve
(AUPEC). It also provides the tools to analyze Individualized
Treatment Rules under budget constraints. Detailed reference in Imai
and Li (2019) <arXiv:1905.05389>.

Michael Lingzhi Li

evalITR

Evaluating Individualized Treatment Rules

Kosuke Imai

Jialu Li

Xiaolong Yang

compute_qoi_user function

<dl><dt>user_itr</dt>
<dd>A user-defined function to create an ITR. The function should take the data as input and return an unit-level continuous score for treatment assignment. We assume those that have score less than 0 should not have treatment. The default is <code>NULL</code>, which means the ITR will be estimated from the <code>estimate_itr</code>.</dd>
<dt>Tcv</dt>
<dd>A vector of the unit-level binary treatment.</dd>
<dt>Ycv</dt>
<dd>A vector of the unit-level continuous outcome.</dd>
<dt>data</dt>
<dd>A data frame containing the variables of interest.</dd>
<dt>ngates</dt>
<dd>The number of gates to be used in the GATE function.</dd>
<dt>budget</dt>
<dd>The maximum percentage of population that can be treated under the budget constraint.</dd>
<dt>...</dt>
<dd>Additional arguments to be passed to the user-defined function.</dd></dl>

Arguments

Compute Quantities of Interest (PAPE, PAPEp, PAPDp, AUPEC, GATE, GATEcv) with user defined functions — compute_qoi_user

<dl>

<dt>user_itr</dt>
<dd>A user-defined function to create an ITR. The function should take the data as input and return an unit-level continuous score for treatment assignment. We assume those that have score less than 0 should not have treatment. The default is <code>NULL</code>, which means the ITR will be estimated from the <code>estimate_itr</code>.</dd>


<dt>Tcv</dt>
<dd>A vector of the unit-level binary treatment.</dd>


<dt>Ycv</dt>
<dd>A vector of the unit-level continuous outcome.</dd>


<dt>data</dt>
<dd>A data frame containing the variables of interest.</dd>


<dt>ngates</dt>
<dd>The number of gates to be used in the GATE function.</dd>


<dt>budget</dt>
<dd>The maximum percentage of population that can be treated under the budget constraint.</dd>


<dt>...</dt>
<dd>Additional arguments to be passed to the user-defined function.</dd>

</dl>

compute_qoi_user: Compute Quantities of Interest (PAPE, PAPEp, PAPDp, AUPEC, GATE, GATEcv) with user defined functions

Description

Usage

Arguments