plotRes {mgm}

Plot summary of resampled sampling distributions

Description

Plots a summary of sampling distributions resampled with the resample() function

Usage

plotRes(object, quantiles = c(.05, .95), labels = NULL, 
        decreasing = TRUE, cut = NULL, cex.label = 0.75, 
        lwd.qtl = 2, cex.mean = 0.55, cex.bg = 2.7, 
        axis.ticks = c(-0.5, -0.25, 0, 0.25, 0.5, 0.75, 1), 
        axis.ticks.mod = NULL, layout.width.labels = .2,
        layout.gap.pw.mod = .15, table = FALSE)

Arguments

object

An output object from the resample() function.

quantiles

A numerical vector of length two, specifying the desired lower/upper quantiles. Defaults to quantiles = c(.05, .95).

labels

A character vector of length p, containing the label of each variable, where p is the number of variables.

decreasing

If TRUE (default), the edges are ordered by the arithmetic mean of the sampling distribution in decreasing order. If FALSE they are ordered in increasing order.

cut

A sequence of integers, specifying which edges are represented. For instance, if decreasing = TRUE and cut = 1:10, summaries for the 10 edges with the largest parameter estimate are displayed. The cut argument can also be used to present the boostrapped CIs in several figures.

cex.label

Text size of the labels.

lwd.qtl

Line width of line indicating the upper/lower quantiles.

cex.mean

Text size of the number indicating the proportion of the estimates whose absolute value is larger than zero.

cex.bg

Size of the white background of the number indicating the proportion of the estimates whose absolute value is larger than zero.

axis.ticks

A numeric vector indicating the axis ticks and labels for the x-axis.

axis.ticks.mod

A numeric vector indicating the axis ticks and labels for the x-axis for moderation effects. If axis.ticks.mod=NULL, the values from axis.ticks for pairwise interactions are used.

layout.width.labels

A positive numeric value which specifies the width of the left-hand-side legend relative to the width of the data panel (or data panels, in case of a moderator model), which have width = 1. Defaults to layout.width.labels = 0.2.

layout.gap.pw.mod

A positive numeric value which specifies the width of the gap between the stability of pairwise effects and moderation effects. Defaults to layout.gap.pw.mod = 0.15.

table

If table = TRUE, the output is presented as a table instead of a figure. Defaults to table = FALSE.

Details

Currently only supports summaries for resampled mgm() objects, and moderated MGMs with a single moderator.

Value

Plots a figure that shows summaries of the resampled sampling distribution for (a set of) all edge parameters. These include the mean, a specified upper and lower quantile and the proportion of parameter estimates whose absolute value is larger than zero.

Author(s)

Jonas Haslbeck <[e-mail]>

See Also

resample(), mgm(), mvar(), tvmgm(), tvmar()

Examples

## Not run: 

# Fit initial model
fit_aut <- mgm(data = as.matrix(autism_data$data),
               type = autism_data$type,
               level = autism_data$lev,
               k = 2)


# Fit bootstrapped models
res_aut <- resample(object = fit_aut, 
                    data = as.matrix(autism_data$data), 
                    nB = 10) # should be more in real applications 

# Plot Summary
plotRes(object = res_aut,
        quantiles = c(.05, .95), 
        labels = NULL, 
        axis.ticks = c(-.25, 0, .25, .5, .75))

  
## End(Not run)