dlso: Latent Structure Optimization Based on Draws

Description

This function provides a partition to summarize a partition distribution using the draws-based latent structure optimization (DLSO) method, which is also known as the least-squares clustering method (Dahl 2006). The method seeks to minimize an estimation criterion by picking the minimizer among the partitions supplied. The implementation currently supports the minimization of several partition estimation criteria. For details on these criteria, see partition.loss.

Usage

dlso(truth, loss = VI(), estimate = NULL)

Value

An integer vector giving the estimated partition, encoded using cluster labels.

Arguments

truth: An integer vector of cluster labels for \(n\) items representing the true clustering. Two items are in the same cluster if their labels are equal. Or, a matrix of \(n\) columns where each row is a clustering.
loss: The loss function to use, as indicated by "binder", "omARI", "VI", "NVI", "ID", "NID", or the result of calling a function with these names. Also supported are "binder.psm", "VI.lb", "omARI.approx", or the result of calling a function with these names, in which case x above can optionally be a pairwise similarity matrix, i.e., \(n\)-by-\(n\) symmetric matrix whose \((i,j)\) element gives the (estimated) probability that items \(i\) and \(j\) are in the same subset (i.e., cluster) of a partition (i.e., clustering).
estimate: An integer vector of cluster labels having the same length as truth representing the estimated clustering. Or, a matrix of \(n\) columns where each row is a clustering.

References

D. B. Dahl (2006), Model-Based Clustering for Expression Data via a Dirichlet Process Mixture Model, in Bayesian Inference for Gene Expression and Proteomics, Kim-Anh Do, Peter Müller, Marina Vannucci (Eds.), Cambridge University Press.

Examples

Run this code

data(iris.clusterings)
dlso(iris.clusterings, loss=VI())
dlso(iris.clusterings, loss=binder())

# Compute expected loss using all draws, but pick the best among the first 10.
dlso(iris.clusterings, loss=VI(), estimate=iris.clusterings[1:10,])