w_get_Hol: Optimal weights for Hölder space

In koohyun-kwon/HTEBand: Uniform confidence band for heterogenous treatment effects

Description

Computes optimal weights and bandwidths for inference for nonparametric regression function values under Hölder space, with an aid of RDHonest package.

Usage

w_get_Hol(
  y,
  x,
  eval,
  C,
  level,
  kern = "triangular",
  se.initial = "EHW",
  se.method = "nn",
  J = 3,
  TE = FALSE,
  d = NULL,
  bw.eq = TRUE
)

Arguments

y	vector of dependent variables
x	vector of regressors
eval	evaluation points
C	bound on the second derivative
level	confidence level
kern	specifies kernel function used in the local regression; default = "triangular". See NPROptBW.fit in RDHonest package for a list of kernels available.
se.initial	method for estimating initial variance for computing optimal bandwidt; default = "EHW". See NPROptBW.fit in RDHonest package for a list of method available.
se.method	methods for estimating standard error of estimate; default = "nn". See NPRreg.fit in RDHonest package for a list of method available.
J	number of nearest neighbors, if "nn" is specified in se.method.
TE	logical specifying whether there are treatment and control groups.
d	a vector of indicator variables specifying treatment and control group status; relevant only when TE = TRUE.
bw.eq	if TRUE, the same bandwidths are used for estimators for treatment and control groups; relevant only when TE = TRUE.

Value

list with components

w.mat

matrix of optimal weigths with a dimension length(y) by length(eval) corresponding to the evaluation points

bw.vec

vector of optimal bandwidths corresponding to the evaluation points

Examples

x <- stats::runif(500, min = -1, max = 1)
y <- x + rnorm(500, 0, 1/4)
eval <- seq(from = -0.9, to = 0.9, length.out = 5)
w_get_Hol(y, x, eval, 1, 0.95)

koohyun-kwon/HTEBand documentation built on Dec. 21, 2021, 7:42 a.m.