jaccard: Jaccard Index

In laderast/Consense: Consense: a package for evaluating multiple clusterings

Description

Given two clusterings C and C', this function calculates the Jaccard Index between the two clusterings. The jaccard index compares the two clusterings by counting and comparing pairs of elements across the two clusterings.

Usage

jaccard(clustering1, clustering2)

Arguments

clustering1	A clustering as defined by the cluster output functions of this package. See ?outputcluster() for more details.
clustering2	same as clustering 1.

Details

Given two clusterings C and C', the Jaccard Index will compare the elements in common between them. This is done by generating an contingency table, whose Nij entry correpsonds to how many elements in common cluster Ci has in common with cluster C'j.

Value

jaccard index

Jaccard index comparing clustering 1 and 2. A value that ranges from 0-1. The more the clusterings are in agreement, the closer the Jaccard index will be to 1.

Author(s)

Ted Laderas ([e-mail])

References

Dudoit, S. and Fridlyand, J. A prediction-based resampling method for estimating the number of clusters in a dataset. Genome Biology. 2002. 3: p. RESEARCH036.

Jain, A.K. and Dubes R.C. Algorithms for Clustering Data. Englewood Cliffs, NJ: Prentice Hall. 1988.

See Also

variationInformation

Examples

  data(choresults)
  clusts <- choresults$clusters
  jaccard(as.data.frame(clusts[[c("UPGMACOR")]]), 
    as.data.frame(clusts[[c("UPGMAEUC")]]))
  

laderast/Consense documentation built on May 20, 2019, 7:32 p.m.