Parametric bootstrap for computing G-best and d-best PCS

Share:

Description

Parametric bootstrap for computing G-best and d-best PCS. This function is called by the wrapper PCS.boot.par.

Usage

1
2
PdofCSGt.bootstrap5(theta, T, D, G, B, SDE, dist = c("normal", "t"), 
	df = 14, trunc = 6, est.names = c("O"))

Arguments

theta

Vector of statistics (or parameters) from which it is desired to select the top t of them

T

Vector of the number of statistics (or parameters) desired to be selected

D

Vector of d-best selection parameters

G

Vector of G-best selection parameters

B

Bootstrap sample size

SDE

Standard error of the statistics theta (row-wise)

dist

Distributional assumption used for estimating PCS

df

Common degrees of freedom for one of the t-statistics in theta; the parameter is only used if dist="t"

trunc

Number of standard errors below the minimum selected population to disregard in the estimation of PCS; it is a truncation parameter to decrease run time

est.names

Kind of shrinkage estimator employed. Default estimator is "O" for the Olkin estimator. Other estimators will be considered for future releases.

Value

An array, the non-empty part of which is a matrix whose rows are the entries of G or D and whose columns are the entries of T. If both G and D are entered, then a list is returned, where the $G element is the G-best matrix, the $d element is the d-best matrix.

Author(s)

Jason Wilson, <jason.wilson@biola.edu>

References

Cui, X. and Wilson, J. 2007. On How to Calculate the Probability of Correct Selection for Large k Populations. University of California, Riverside Statistics Department Technical Report 297. http://www.bubbs.biola.edu/~jason.wilson/Article2_tech_techreport.pdf

See Also

PCS.boot.par