Cluster analysis fit statistic using bootstrap distributions

Note: This will always use bootstrapping.

Usage

cluster_overlap_fit(df, k = 2:6, verbose = interactive(), ...)

# S3 method for class 'cluster.overlap'
plot(x, se_factor = 1, ...)

Arguments

df: data frame containing the variables to use for cluster.
k: an integer vector for the cluster sizes to estimate the overlap for.
verbose: whether to print status as the boostrap samples are estimated.
...: currently not used.
x: results from cluster_overlap_fit().
se_factor: factor to multiple the standard error. For example, for a 95% confidence interval set se_factor = 1.96.

Value

an object that inherits from a data.frame with the following columns:

variable: the variable name.
C1: the label for cluster 1.
C2: the label for cluster 2.
overlap: how much of the bootstrap distributions overlap from the distribution of cluster centers for clusters 1 and 2 for variable.
k: the number of clusters.

There is a custom plot() function for the returned object.

a ggplot2 expression.

Examples

if (FALSE) { # \dontrun{
data(daacs, package='clav')
cluster_vars <- c("Motivation", "Metacognition", "Strategies", "Mathematics", "Reading", "Writing")
cof <- cluster_overlap_fit(daacs[,cluster_vars])
plot(cof)
} # }