HMMVBclust-class: Class "HMMVBclust" to represent clustering results with...

In HDclust: Clustering High Dimensional Data with Hidden Markov Model on Variable Blocks

Description

An S4 class to represent a clustering result based on HMM-VB. New instances of the class are created by hmmvbClust.

Methods

• show signature(object = "HMMVBclust") : show clustering results based on HMM-VB.

• plot signature(x = "HMMVBclust", y = "missing", method = "t-sne", ...) : plot clustering results. 'method' controls the visualization algorithm. Two algorithms are supported: method = 'PCA' plots the data using 2 component PCA space; and method = 't-SNE' plots the data using 2 component t-SNE space. Default setting is t-SNE.

• getClustParam signature(object = "HMMVBclust") : accessor for 'clustParam' slot.

• getLoglikehd signature(object = "HMMVBclust") : accessor for 'Loglikehd' slot.

• getClsid signature(object = "HMMVBclust") : accessor for 'clsid' slot.

• getSize signature(object = "HMMVBclust") : accessor for 'size' slot.

Slots

data

The input data matrix

clustParam

A list with cluster parameters:

ncls

The number of clusters (same as the number of modes)

mode

A numeric matrix with cluster modes. kth row of the matrix stores coordinates of the kth mode.

ndseq

The number of distinct Viterbi sequences for the dataset

vseqid

An integer vector representing the map between Viterbi sequences and clusters. kth value in the vector stores cluster id for kth Viterbi sequence.

vseq

A list with integer vectors representing distinct Viterbi sequences for the dataset

sigma

A numeric vector with the dataset variance

clsid

An integer vector with cluster ids.

Loglikehd

Loglikelihood value for each data point.

size

An integer vector with cluster sizes.

HDclust documentation built on May 2, 2019, 9:20 a.m.