voiClusters: Validity-Relevant Clusters

In funfield-lab/fancyr: Fancy Statistics for Correlational (r) Analyses

Validity-Relevant Clusters

Description

This automates a set of procedures for finding clusters of items that relate in similar ways to a variable of interest.

This should produce a set of factors that relates somewhat independently to the variable of interest

Usage

voiClusters(x, rvoi, threshold = 0.15, nk = 6, cname = "c", weight = "unit")

Arguments

x	item set to be associated with the VOI. It is suggested that items are placed on a [-1,1] range pprior to analysis to decrease the variability in cluster scores produced by missing values (this can be done using the cx function)
rvoi	items' correlation with the 'variable of interest' (these can be partial correlations)
threshold	minimum correlation to be included in the 'set of relevant items'
nk	the number of clusters you ultimately wish to extract
cname	name of the cluster (defaults to 'c')
weight	should items receive unit-weight in clusters (after clearing threshold), or should they be weighted by level of association with the VOI?

Value

cluster scores, cluster membership, and full dendrogram

funfield-lab/fancyr documentation built on Nov. 21, 2023, 2:42 p.m.