nomclust (version 1.1.1106)

nomprox: Nominal Clustering based on a Proximity Matrix

Description

Based on the original dataset and the proximity matrix, the function computes cluster membership variables for a user-defined number of cluster solutions. Optionally, it evaluates clustering results using six evaluation criteria based on the within-cluster variability: Within-cluster mutability coefficient (WCM), Within-cluster entropy coefficient (WCE), Pseudo tau coefficient (PSTau), Pseudo uncertainty coefficient (PSU) and Pseudo F, Indices based on the mutability (PSFM) and the entropy (PSFE).

Usage

nomprox(data, prox_matrix, clu_low = 2, clu_high = 6, eval = TRUE,
  method = "complete")

Arguments

data

data frame or a matrix with cases in rows and variables in colums. Cases are characterized by nominal (categorical) variables coded as numbers.

prox_matrix

full proximity matrix computed using any similarity measure from the data analyzed.

clu_low

numeric value expressing the lower bound for number of cluster solutions.

clu_high

numeric value expressing the higher bound for number of cluster solutions.

eval

logical operator; if TRUE, there is performed an evaluation of clustering results

method

character string defining the clustering method. The following methods can be used: "average", "complete", "single".

Value

Function returns a data frame, where the rows express a serie of cluster solutions and columns clustering evaluation statistics in a following order: WCM, WCE, PSTau, PSU, PSFM, PSFE.

See Also

nomclust, evalclust.

Examples

# NOT RUN {
#sample data
data(data20)
#computation of a proximity matrix using the iof similarity measure
matrix <- iof(data20)
#creation of a dataset with cluster memberships
hca <- nomprox(data20, matrix, clu_high = 5, method = "complete")
#getting evaluation statistics
eval <- hca$eval
#getting cluster membership variables
mem <- hca$mem 

# }