getHillBootstrap {braidrm}

Bootstrap Hill Equation Parameter Confidence Intervals

Description

Uses sampling residuals with replacement to bootstrap a distribution of Hill equation parameters for the construction of confidence intervals on those parameters.

Usage

getHillBootstrap(hfitl, ciLevs = c(0.025,0.975), mi = NULL)

Arguments

hfitl

a list of Hill equation model fits produced by findBestHill

ciLevs

a two-element vector specifying the lower and upper percentiles of the desired confidence intervals. Default is c(0.025,0.975), which yields a 95% confidence interval.

mi

index of the desired model to bootstrap and construct confidence intervals on. If NULL (the default), the value of bestModIdx in hfitl will be used.

Value

A list of the same form as produced by findBestHill, with the following additional components:

ciPass

TRUE if at least 50% of bootstrapping trials result in a converged fit

ciLevs

Equal to the input parameter ciLevs

ciMInd

Index of the model used to construct the bootstrapped confidence interval

ciVec

Vector of lower and upper bounds of confidence intervals on all freely varying parameters, in order

bCoefs

Four column array containing complete Hill equation parameters for all resampled fits

Author(s)

Nathaniel R. Twarog

See Also

findBestHill, evalHillEqn

Examples

data(es8olatmz)
drv <- es8olatmz$compound2=="DMSO"
hfit <- findBestHill(act~conc1,es8olatmz[drv,],defaults=c(0,-2.7))
hfit1 <- getHillBootstrap(hfit)

# Setting 'mi' to 4 constructs a confidence interval on the four-
# parameter Hill model, overriding the 'best' model index
hfit2 <- getHillBootstrap(hfit,ciLevs=c(0.05,0.95),mi=4)