specContrasts: Compute specificities (importances) and contrasts ("partial"...

In clamix: Clustering modal multi valued symbolic data

Description

Computes specificities for a variable comparing two cluster leaders and contrasts (partial importances) of the second cluster leader towards first for each variable component.

Usage

specContrasts(rs,rc,j)

Arguments

rs	first cluster leader in the form of symObject, usually leader of whole dataset.
rc	second cluster leader in the form of symObject.
j	an index (or a name) of a variable.

Value

A list of

specificity	A value.
contrasts	A vector of components of NaN, Inf (see reference) or a value x where |x| > 1.

Author(s)

Natasa Kejzar

References

Batagelj,~V., Kejzar,~N. and Korenjak-Cerne,~S., Clustering of Modal Valued Data, a chapter in Analysis of Distributional Data (Ed. Brito,~P.), (submitted)

See Also

hclust, hclustSO, leaderSO, print.symObject, empty.symObject

Examples

data(TIMSSSlo)
tsdata = create.symData(TIMSSSlo)
dendro <- hclustSO(tsdata)
rs = dendro$leaders[[nrow(dendro$merge)]] # the whole dataset
rc = dendro$leaders[[84]] # first cluster from reference chapter

specContrasts(rs,rc,1) # first variable ("MT2GTAUT") 

clamix documentation built on May 2, 2019, 4:51 p.m.