Function to compute the prediction strength of a clustering model

Description

This function computes the prediction strength of a clustering model as published in R. Tibshirani and G. Walther 2005.

Usage

1
ps.cluster(cl.tr, cl.ts, na.rm = FALSE)

Arguments

cl.tr

Clusters membership as defined by the original clustering model, i.e. the one that was not fitted on the dataset of interest.

cl.ts

Clusters membership as defined by the clustering model fitted on the dataset of interest.

na.rm

TRUE if missing values should be removed, FALSE otherwise.

Value

ps

the overall prediction strength (minimum of the prediction strengths at cluster level).

ps.cluster

Prediction strength for each cluster

ps.individual

Prediction strength for each sample.

Author(s)

Benjamin Haibe-Kains

References

R. Tibshirani and G. Walther (2005) "Cluster Validation by Prediction Strength", Journal of Computational and Graphical Statistics, 14(3):511–528.

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
## load SSP signature published in Sorlie et al. 2003
data(ssp2003)
## load NKI data
data(nkis)
## SP2003 fitted on NKI
ssp2003.2nkis <- intrinsic.cluster(data=data.nkis, annot=annot.nkis,
  do.mapping=TRUE, std="robust",
  intrinsicg=ssp2003$centroids.map[ ,c("probe", "EntrezGene.ID")],
  number.cluster=5, mins=5, method.cor="spearman",
  method.centroids="mean", verbose=TRUE)
## SP2003 published in Sorlie et al 2003 and applied in VDX
ssp2003.nkis <- intrinsic.cluster.predict(sbt.model=ssp2003,
  data=data.nkis, annot=annot.nkis, do.mapping=TRUE, verbose=TRUE)
## prediction strength of sp2003 clustering model
ps.cluster(cl.tr=ssp2003.2nkis$subtype, cl.ts=ssp2003.nkis$subtype,
  na.rm = FALSE)

Want to suggest features or report bugs for rdrr.io? Use the GitHub issue tracker.