SnoopingCVNearBd: Snooping-adjusted critical values for estimation near a...

In kolesarm/BWSnooping: Confidence invervals robust to bandwidth snooping

Description

Calculate one- and two-sided critical values c_{1-α}(t;k) for values of t in bwratios based on evaluating the Gaussian process \hat{\mathbb{H}}(h) at ngr values of h in the interval [1/t,1].

Usage

SnoopingCVNearBd(S, T, bwratios, kernel, order, db, ngr, alpha = c(0.1, 0.05,
  0.01))

Arguments

S	number of draws of the Gaussian process \hat{\mathbb{H}}(h)
T	number of draws from a normal distribution in each draw of the Gaussian process
bwratios	Bandwidth ratios of maximum to minimum bandwidth for which to compute critical values
kernel	Kernel function k(u) supported on [-1,1] that takes a vector or a matrix as an argument u.
order	Order of local linear regression
db	Local distance to boundary, equal to x_{0}/\underline{h}, where x_{0} is point of interest.
ngr	number of grid points at which to evaluate the Gaussian process
alpha	A vector of values determining the confidence level 1-α at which to compute critical values

kolesarm/BWSnooping documentation built on May 20, 2019, 12:54 p.m.