clvpredictStrength: cluster validation by prediction strength

In nguforche/UnsupRF: Unsupervised Random Forest Clustering

Description

Calculate the mediods of a clustering by finding the point that has the minimum average/sum distance to all other points in the cluster. Proximity between data points can be provided by RFdist or using the Gower's general similarity coefficient.

Usage

clvpredictStrength(dat, ...)

## Default S3 method:
clvpredictStrength(dat, dist.mat, method = "ward.D2",
  nBoots = 20, krange = 2:5, balanced = FALSE, parallel = FALSE,
  mc.cores = 2, OOB = TRUE, seed = 12345, ...)

## S3 method for class 'clvpredictStrength'
plot(x, perf.measure = "Sensitivity")

## S3 method for class 'clvpredictStrength'
print(x, ...)

Arguments

dat	a data matrix
...	further arguments passed to or from other methods.
dist.mat	dissimilarity matrix obtained from the data matrix "dat"
method	hclust clustering method
nBoots	number of bootstraps
krange	integer vector. Numbers of clusters which are to be tried
parallel	run in parallel ?
mc.cores	number of CPU cores
OOB	use out-of-bag from bootstrap as test data?
seed	random seed
balance	perform balance bootstrap ?

Details

This function is experimental .. more details coming

nguforche/UnsupRF documentation built on May 5, 2019, 4:51 p.m.