IPS_exp: Integrated Propensity Score estimator based on exponential...

In pedrohcgs/IPS: IPS: Covariate Distribution Balance via Integrated Propensity Scores

Description

Integrated Propensity Score estimator based on exponential weighting function

Usage

IPS_exp(
  d,
  x,
  xbal = NULL,
  X.trans = "normal",
  Treated = FALSE,
  beta.initial = NULL,
  lin.rep = TRUE,
  whs = NULL,
  x_keep = FALSE,
  maxit = 50000
)

Arguments

d	An n x 1 vector of binary treatment adoption indicators.
x	An n x k matrix of covariates to be used in the propensity score. First element must be a vector of 1's.
xbal	An n x l, l≤q k, matrix of “raw” covariares to be balanced (does not need to include interaction terms). Default is NULL, which will use the same as x.
X.trans	description of which transformation of covariates is used to enforce compactness. The alternatives are 'normal' (default), and 'arctan'.
Treated	Default is FALSE, which aims to achieve covariate distribution balance among treated, untreated and overall subpopulations. If TRUE, then the estimator aims to achieve covariate distribution balance for the treated subpopulation.
beta.initial	An optional k x 1 vector of initial values for the parameters to be optimized over.
lin.rep	Logical argument to whether an estimator for the asymptotic linear representation of the IPS parameters should be provided. Deafault is TRUE.
whs	An optional n x 1 vector of weights to be used. If NULL, then every observation has the same weights.
x_keep	Default is FALSE. If TRUE, we return covariate matrix in the output.
maxit	The maximum number of iterations. Defaults to 50000. = FALSE). Default is 999 if boot = TRUE

Value

A list containing the following components:

coefficients	The estimated IPS_exp coefficients
fitted.values	The IPS_exp fitted probabilities
linear.predictors	The IPS_exp estimated index (X'beta)
lin.rep	An estimator of the IPS_exp coefficients' asymptotic linear representation
converged	An integer code. 0 indicates successful completion
x	The model matrix (i.e. the matrix of covariates used to estimate the IPS_exp parameters). Only returned if x_keep = TRUE.

References

Sant'Anna, Pedro H. C, Song, Xiaojun, and Xu, Qi (2019), Covariate Distribution Balance via Propensity Scores, Working Paper <https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3258551>.

pedrohcgs/IPS documentation built on Dec. 22, 2021, 7:39 a.m.