Become an expert in R — Interactive courses, Cheat Sheets, certificates and more!
Get Started for Free

clustModes

Hierarchical clustering of density modes

Description

This function performs hierarchical clustering of density modes found by hmmvbFindModes().

Usage

clustModes(modes, cutree.args, hclust.args = NULL, dist.args = NULL)

Arguments

modes

An object of class 'HMMVBclust' returned by hmmvbFindModes().
cutree.args

A list with arguments to cutree method from package stats.
hclust.args

A list with arguments to hclust method from package stats. It is used to specify a linkage method for hierarchical clustering of density modes. Complete linkage is used by default. See hclust for details.
dist.args

A list with arguments to dist method from package stats. It is used to specify a distance metric for the density modes. Euclidian distance is used by default. See dist for details.

Value

An object of class 'HMMVBclust' with new cluster labels and cluster sizes. Note that coordinates of modes after merging are not calculated and clustParam field is empty.

See Also

Examples

Vb <- vb(1, dim=4, numst=2)
set.seed(12345)
hmmvb <- hmmvbTrain(unique(iris[,1:4]), VbStructure=Vb)
modes <- hmmvbFindModes(unique(iris[,1:4]), model=hmmvb)

# default mode clustering
merged <- clustModes(modes, cutree.args=list(h=1.0))

# mode clustering using Manhattan distance
merged <- clustModes(modes, dist.args=list(method="manhattan"), cutree.args=list(h=1.0))

# mode clustering using single linkage
merged <- clustModes(modes, hclust.args=list(method="single"), cutree.args=list(h=1.0))
HDclust

Clustering High Dimensional Data with Hidden Markov Model on Variable Blocks

v1.0.3
GPL (>= 2)
Authors
Yevhen Tupikov [aut], Lin Lin [aut], Lixiang Zhang [aut], Jia Li [aut, cre]
Initial release
2019-04-05

We don't support your browser anymore

Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.