summary.salso.estimate: Summary of Partitions Estimated Using Posterior Expected Loss

In salso: Search Algorithms and Loss Functions for Bayesian Clustering

Summary of Partitions Estimated Using Posterior Expected Loss

Description

Assessing the quality of clusters in a partition estimate is added by this function. The result can then be plotted with plot.salso.summary. The current implementation of the calculation of these summaries is not terribly efficient and may be improved in the future.

Usage

## S3 method for class 'salso.estimate'
summary(object, alternative, orderingMethod = 1, ...)

Arguments

object	An object returned by the salso function.
alternative	An optional argument specifying an alternative clustering to use instead of that provided by object. Use this feature to obtain numerical and graphical summaries of a clustering estimate from other procedures. This clustering must be provided in canonical form: cluster labels as integers starting at 1 for the first observation and incrementing by one for each new label.
orderingMethod	An integer giving method to use to order the observations for a heatmap plot. Currently values 1 or 2 are supported.
...	Currently ignored.

Value

A list containing the estimate, the pairwise similarity matrix, the mean pairwise similarity matrix, the score and mean pairwise similarity for each observation, exemplar observation for each cluster, a dendrogram object, a vector for ordering observations in the heatmap plot, the size of each cluster, and the number of clusters.

Examples

# For examples, use 'nCores=1' per CRAN rules, but in practice omit this.
data(iris.clusterings)
draws <- iris.clusterings
est <- salso(draws, nCores=1)
summ <- summary(est)
plot(summ, type="heatmap")
plot(summ, type="mds")
plot(summ, type="pairs", data=iris)
plot(summ, type="dendrogram")

salso documentation built on Aug. 29, 2025, 5:26 p.m.