Skip to contents

Composite-Weight Stability Under Case Resampling

Source: R/mcml_pc.R
loading_stability.Rd

Experimental. Bootstraps the item weights of a build_mcml_pc fit: rows of the raw data are resampled, the node-level network is re-estimated each time, and the connectivity-based composite weights are recomputed. Wide intervals mean the weighting (and therefore the "loadings" macro network) should not be over-interpreted.

Usage

loading_stability(x, iter = 200L, ci_level = 0.05, seed = NULL)

# S3 method for class 'pc_loading_stability'
print(x, digits = 3, ...)

# S3 method for class 'pc_loading_stability'
plot(x, ...)

Arguments

x

An mcml_pc object that carries raw data. For the print() and plot() methods: an object of class pc_loading_stability.

iter

Integer. Bootstrap replicates (default 200; node-level re-estimation makes this heavier than a plain bootstrap).

ci_level

Numeric. Significance level for percentile CIs (default 0.05).

seed

Integer or NULL. RNG seed.

digits

Number of digits to display (default 3).

...

In plot.pc_loading_stability() and print.pc_loading_stability(): Additional arguments (ignored).

Value

An object of class "pc_loading_stability": a list with summary (tidy data frame: node, cluster, weight, boot_mean, boot_sd, ci_lower, ci_upper, sign_flips - the proportion of replicates in which the item's sign differed from the observed one), boot_weights (iter x n_nodes matrix), iter, and ci_level. Has print and plot methods.

In print.pc_loading_stability(): x, invisibly.

In plot.pc_loading_stability(): A ggplot object.

Methods

  • plot.pc_loading_stability(): Signed composite weights with bootstrap percentile intervals, faceted by cluster.

Examples

set.seed(1)
f <- stats::rnorm(100)
g <- stats::rnorm(100)
df <- data.frame(a1 = f + stats::rnorm(100), a2 = f + stats::rnorm(100),
                 a3 = f + stats::rnorm(100), b1 = g + stats::rnorm(100),
                 b2 = g + stats::rnorm(100), b3 = g + stats::rnorm(100))
cl <- list(A = c("a1", "a2", "a3"), B = c("b1", "b2", "b3"))
fit <- build_mcml_pc(df, cl, aggregation = "loadings",
                     method = "cor")
stability <- loading_stability(fit, iter = 50, seed = 1)
stability
#> Composite-Weight Stability (case bootstrap, experimental)
#>   50 replicates | 95% percentile CIs
#> 
#>  node cluster weight boot_mean boot_sd ci_lower ci_upper sign_flips
#>    a1       A  0.338     0.332   0.021    0.292    0.370          0
#>    a2       A  0.336     0.339   0.018    0.308    0.375          0
#>    a3       A  0.326     0.329   0.021    0.290    0.367          0
#>    b1       B  0.316     0.317   0.022    0.276    0.348          0
#>    b2       B  0.350     0.346   0.022    0.310    0.393          0
#>    b3       B  0.334     0.337   0.015    0.307    0.360          0