twostageTE-package: Threshold value estimation using two-stage plans

In twostageTE: Two-Stage Threshold Estimation

Description

This package implements a variety of nonparametric methods for computing one-stage and two-stage confidence intervals and point estimates of threshold values.

Details

Package:	twostageTE
Type:	Package
Version:	1.0
Date:	2013-05-23
License:	GPL-2

The user interacts with the package by utilizing two functions: stageOneAnalysis and stageTwoAnalysis. These functions take the sampled explanatory variable and corresponding responses at the first and second stage, respectively, and outputs point estimate and confidence intervals based on different user specific procedures.

Author(s)

Shawn Mankad Maintainer: Shawn Mankad <[e-mail]>

References

Shawn Mankad, George Michailidis, Moulinath Banerjee (2015). Threshold Value Estimation Using Adaptive Two-Stage Plans in R. Journal of Statistical Software, 67(3), 1-19. doi:10.18637/jss.v067.i03

Examples

## Simulating the (wiggly) isotonic Sine function ##
sampleData=function(n, lower, upper) {
x=runif(n, lower, upper)
y=(1/40)*sin(6*pi*x) + 1/4 + x/2 + (1/4)*x^2
+ rnorm(n=length(x), sd=0.1)
return(list(X=x, Y=y))
}
Budget=100
d0=0.5
threshold = (1/40)*sin(6*pi*d0) + 1/4 + d0/2 + (1/4)*d0^2

n1=floor(Budget*0.25)
n2=Budget - n1
samp = sampleData(n1, lower=0, upper=1)
X = samp$X
Y = samp$Y
## Two Stage IR+IR ##
stageOne_IR=stageOneAnalysis(X, Y, threshold, type="IR-wald", 0.99)
samp2 = sampleData(n2, lower=stageOne_IR$L1, upper=stageOne_IR$U1)
X2 = samp2$X
Y2 = samp2$Y
twoStageIR = stageTwoAnalysis(stageOne_IR, X2, Y2, type="IR-wald", 0.95)
## Two Stage LR+LR ##
stageOne_LR=stageOneAnalysis(X, Y, threshold, type="IR-likelihood", 0.99)
samp2 = sampleData(n2, lower=stageOne_LR$L1, upper=stageOne_LR$U1)
X2 = samp2$X
Y2 = samp2$Y
twoStageLR = stageTwoAnalysis(stageOne_LR, X2, Y2, 
    type="IR-likelihood", 0.95)
## Two Stage IR+Local Linear ##
X2 = c(rep(stageOne_IR$L1,37),rep(stageOne_IR$U1,38))
Y2=X2^2+rnorm(n=length(X2), sd=0.1)
twoStageLinear=stageTwoAnalysis(stageOne_IR, explanatory = X2, response = Y2,
    type = "locLinear", level = 0.95)

twostageTE documentation built on May 1, 2019, 9:18 p.m.