bootstrap_critical_cutoffs: Bootstrap critical segmentation magnitude

Description

Bootstraps critical segmentation magnitude values for a segmag object under the null hypothesis that all key presses were randomly distributed (uniformly) across the experiment (time_min to time_max).

Usage

bootstrap_critical_cutoffs(segmag, n_bootstrap, critical_probs, segmag_substract = NULL, visualize = FALSE, save_as = NULL)

Arguments

segmag

object of class segmag

n_bootstrap

numeric, number of bootstrap iterations

critical_probs

numeric vector of probabilities, e.g. c(.95,.99)

segmag_substract

object of class segmag. If this value is set than keypress times in segmag and segmag_substract are both randomized and the critical_cutoffs relate to the difference of the segmentation magnitude in segmag minus the segmentation magnitude in segmag_substract

visualize

logical, visualize ordered maxima (and minima if segmag_substract is set) of bootstrapping iterations (Note: Enabling this option might require a lot of RAM with large data sets or large values of n_bootstrap)

save_as

character, filename where to save raw bootstrapping data and plot (optional)

Value

critical segmentation magnitudes; If segmag_substract is NULL, then the return value is a numeric vector. Otherwise a list with critical maxima cutoffs and critical minima cutoffs is returned.

Details

During each bootstrapping iteration, the key presses are randomly distributed (drawn from uniform distribution ranging from time_min to time_max). Then, segmentation magnitude is calculated with those random key press times (note that ids are retained, that is each participant "makes" the same amount of key presses as in the original experiment). The local maxima in segmentation magnitude resulting from the random key press times are ordered according to their size. The largest maximum is kept.

The function returns the critical_probs quantiles of the vector of those largest maxima obtained across n_bootstap iterations

This function can also be used to bootstrap the critical maxima and minima cutoffs of a difference function of two segmag objects. To do so, segmag and segmag_substract must be defined. All values will be related to the difference of segmag - segmag_substract (Keypress times in segmag and segmag_substract are randomized independently).

Examples

Run this code

#see ?segmag for an example