honest_splitting

Splits data into discovery and inference sub-samples.

internal

Provides a new method for interpretable heterogeneous
treatment effects characterization in terms of decision rules
via an extensive exploration of heterogeneity patterns by an
ensemble-of-trees approach, enforcing high stability in the
discovery. It relies on a two-stage pseudo-outcome regression, and
it is supported by theoretical convergence guarantees. Bargagli-Stoffi,
F. J., Cadei, R., Lee, K., & Dominici, F. (2023) Causal rule ensemble:
Interpretable Discovery and Inference of Heterogeneous Treatment Effects.
arXiv preprint <doi:10.48550/arXiv.2009.09036>.

Falco Joannes Bargagli Stoffi

Interpretable Discovery and Inference of Heterogeneous Treatment
Effects

Naeem Khoshnevis

Daniela Maria Garcia

Riccardo Cadei

Kwonsang Lee

honest_splitting function

<dl><dt>y</dt>
<dd>An observed response vector.</dd>
<dt>z</dt>
<dd>A treatment vector.</dd>
<dt>X</dt>
<dd>A features matrix.</dd>
<dt>ratio_dis</dt>
<dd>A double number indicating the ratio of data delegated to
the discovery sub-sample.</dd>
<dt>ite</dt>
<dd>A vector of estimated ITE.</dd></dl>

Arguments

Honest splitting — honest_splitting

<dl>

<dt>y</dt>
<dd>An observed response vector.</dd>


<dt>z</dt>
<dd>A treatment vector.</dd>


<dt>X</dt>
<dd>A features matrix.</dd>


<dt>ratio_dis</dt>
<dd>A double number indicating the ratio of data delegated to
the discovery sub-sample.</dd>


<dt>ite</dt>
<dd>A vector of estimated ITE.</dd>

</dl>

honest_splitting: Honest splitting

Description

Usage

Value

Arguments