boot.cond.mean.per: Percentile Bootstrap Conditional Mean CI in Simple Linear...

In wBoot: Bootstrap Methods

Description

Determines a confidence interval for a conditional mean in simple linear regression, using the percentile bootstrap method.

Usage

boot.cond.mean.per(x, y, xp, conf.level = 0.95, R = 9999)

Arguments

x	a (nonempty) numeric vector of predictor-variable data values.
y	the corresponding numeric vector of response-variable data values.
xp	the value of the predictor variable for which to find the CI for the conditional mean of the response variable.
conf.level	the confidence level (between 0 and 1); default is 0.95.
R	the number of bootstrap replications; default is 9999.

Value

A list with class "boot.regcor" containing the following components:

Boot.values	the point estimates (fits) obtained from the bootstrap.
Confidence.limits	the upper and lower limits of the confidence interval.
Header	the main title for the output.
Variable.1	the predictor variable.
Variable.2	the response variable.
n	the sample size.
Statistic	the name of the statistic, here fit.
Observed	the observed point estimate (fit).
Replications	the number of bootstrap replications.
Mean	the mean of the bootstrap values.
SE	the standard deviation of the bootstrap values.
Bias	the difference between the mean of the bootstrap values and the observed value.
Percent.bias	the percentage bias: 100*|Bias/Observed|.
Null	always NULL for this function.
Alternative	always NULL for this function.
P.value	always NULL for this function.
p.value	always NULL for this function.
Level	the confidence level.
Type	always NULL for this function.
Confidence.interval	the confidence interval.
cor.ana	a logical; always FALSE for this function.

Warning

This routine should be used only when bias is small and the sampling distribution is roughly symmetric, as indicated by the output of the bootstrap. Otherwise, use the BCa version.

Author(s)

Neil A. Weiss

Examples

# Lot size, house size, and value for a sample of homes in a particular area.
data("homes")
str(homes)
attach(homes)

# 95% (default) CI for the conditional mean value of a 3000 sq.ft. home,
# with 999 bootstrap replications.
boot.cond.mean.per(HOUSE.SIZE, VALUE, 3000, R = 999)

# 90% CI for the conditional mean value of a 3000 sq.ft. home, with
# 999 bootstrap replications.
boot.cond.mean.per(HOUSE.SIZE, VALUE, 3000, conf.level = 0.90, R = 999)

detach(homes) # clean up

Example output

Loading required package: boot
Loading required package: simpleboot
Simple Bootstrap Routines (1.1-7)
'data.frame':	44 obs. of  3 variables:
 $ HOUSE.SIZE: num  2311 2968 3773 1934 5466 ...
 $ LOT.SIZE  : num  2.37 2.09 2.21 2.21 2.1 2.06 2.03 2.44 2.14 2.63 ...
 $ VALUE     : num  396 355 586 254 646 278 279 748 546 338 ...


 RESULTS OF PERCENTILE BOOTSTRAP FOR CONDITIONAL MEAN CORRESPONDING TO HOUSE.SIZE = 3000 

    SUMMARY  Predictor Response  n Statistic Observed
 STATISTICS HOUSE.SIZE    VALUE 44       Fit 446.3234

 BOOTSTRAP Replications     Mean       SE Bias Percent.bias
   SUMMARY          999 447.4446 15.32613 1.12        0.251

 CONFIDENCE Level Confidence.interval
   INTERVAL   95%      (419.3, 476.6)




 RESULTS OF PERCENTILE BOOTSTRAP FOR CONDITIONAL MEAN CORRESPONDING TO HOUSE.SIZE = 3000 

    SUMMARY  Predictor Response  n Statistic Observed
 STATISTICS HOUSE.SIZE    VALUE 44       Fit 446.3234

 BOOTSTRAP Replications     Mean       SE  Bias Percent.bias
   SUMMARY          999 446.4453 14.79349 0.122       0.0273

 CONFIDENCE Level Confidence.interval
   INTERVAL   90%      (422.9, 472.1)

wBoot documentation built on May 1, 2019, 7:31 p.m.