itemStability: Item Stability Statistics from `bootEGA`

Description

Based on the bootEGA results, this function computes and plots the number of times an item (variable) is estimated in the same factor/dimension as originaly estimated by EGA (item.replication). The output also contains each item's replication frequency (i.e., proprotion of bootstraps that an item appeared in each dimension; item.dim.rep) as well as the average network loading for each item in each dimension (item.loadings).

Usage

itemStability(bootega.obj, orig.wc, item.freq = 0.1,
  plot.item.rep = TRUE)

Arguments

bootega.obj

A bootEGA object

orig.wc

Numeric or character. A vector with community numbers or labels for each item. Typically uses community results (wc) from EGA

item.freq

A value for lowest frequency allowed in item.dim.rep output. Removes noise from table to allow for easier interpretation. Defaults to .10

plot.item.rep

Should the plot be produced for item.replication? If TRUE, then a plot for the item.replication output will be produced.#' Defaults to TRUE

Value

Returns a list containing:

item.replication

The proporton of times each item replicated within the defined dimension

item.dim.rep

The proportion of times each item replicated within each possible dimension. Dimensions greater than the maximum number used in the orig.wc argument are labeled based on the largest remaining components after the dimensions used to orig.wc

item.loadings

Matrix of the average standardied network loading (computed using net.loads) for each item in each dimension

A matrix containing the community membership values for each bootstrapped sample. The values correspond to the values input for the orig.wc argument

References

Danon, L., Diaz-Guilera, A., Duch, J., & Arenas, A. (2005). Comparing community structure identification. Journal of Statistical Mechanics: Theory and Experiment, 9, P09008. <doi:10.1088/1742-5468/2005/09/P09008>

Examples

Run this code

# NOT RUN {
# Load data
wmt <- wmt2[,7:24]

# }
# NOT RUN {
# Estimate EGA network
ega.wmt <- EGA(data = wmt, model = "glasso")

# Estimate dimension stability
boot.wmt <- bootEGA(data = wmt, n = 100, typicalStructure = TRUE,
plot.typicalStructure = TRUE, model = "glasso",
type = "parametric", ncores = 4)

# }
# NOT RUN {
# Estimate item stability statistics
itemStability(boot.wmt, orig.wc = ega.wmt$wc)

# }