bootstrap.analysis: Performs a Bootstrap with Blocking Analysis of a Timeseries

View source: R/bootstrapnumber.R

bootstrap.analysisR Documentation

Performs a Bootstrap with Blocking Analysis of a Timeseries

Description

Performs a Bootstrap with Blocking Analysis of a Timeseries

Usage

bootstrap.analysis(data, skip = 0, boot.R = 100, tsboot.sim = "geom",
  pl = FALSE, boot.l = 2)

Arguments

data

a numerical vector containing the time series

skip

integer value providing the warm up phase length.

boot.R

number of bootstrap samples. See also boot, and tsboot.

tsboot.sim

the sim parameter of tsboot.

pl

logical, indicating whether or not to plot the result.

boot.l

block length for blocked bootstrap.

Details

the routine will compute the error, the error of the error and the integrated autocorrelation time for different block size using a bootstrap analysis. The blocksize is systematically increased starting from 1 until (length(data)-skip)/blocksize < 20. Note that only data is kept in exact multiples of the block length.

Value

returns a data frame containing the mean value, the error approximation, the estimate of the error of the error, the value of tau int and the bias for all block sizes.

Author(s)

Carsten Urbach, carsten.urbach@liverpool.ac.uk

See Also

for an alternative way to analyse such time series see uwerr and computeacf

Examples


data(plaq.sample)
plaq.boot <- bootstrap.analysis(plaq.sample, pl=TRUE)


hadron documentation built on Sept. 9, 2022, 5:06 p.m.