assign_cluster: Assign Assessors to Clusters

In BayesMallows: Bayesian Preference Learning with the Mallows Rank Model

Assign Assessors to Clusters

Description

Assign assessors to clusters by finding the cluster with highest posterior probability.

Usage

assign_cluster(model_fit, soft = TRUE, expand = FALSE)

Arguments

model_fit	An object of type BayesMallows, returned from compute_mallows().
soft	A logical specifying whether to perform soft or hard clustering. If soft=TRUE, all cluster probabilities are returned, whereas if soft=FALSE, only the maximum a posterior (MAP) cluster probability is returned, per assessor. In the case of a tie between two or more cluster assignments, a random cluster is taken as MAP estimate.
expand	A logical specifying whether or not to expand the rowset of each assessor to also include clusters for which the assessor has 0 a posterior assignment probability. Only used when soft = TRUE. Defaults to FALSE.

Value

A dataframe. If soft = FALSE, it has one row per assessor, and columns assessor, probability and map_cluster. If soft = TRUE, it has n_cluster rows per assessor, and the additional column cluster.

See Also

Other posterior quantities: compute_consensus(), compute_posterior_intervals(), get_acceptance_ratios(), heat_plot(), plot.BayesMallows(), plot.SMCMallows(), plot_elbow(), plot_top_k(), predict_top_k(), print.BayesMallows()

Examples

# Fit a model with three clusters to the simulated example data
set.seed(1)
mixture_model <- compute_mallows(
  data = setup_rank_data(cluster_data),
  model_options = set_model_options(n_clusters = 3),
  compute_options = set_compute_options(nmc = 5000, burnin = 1000)
)

head(assign_cluster(mixture_model))
head(assign_cluster(mixture_model, soft = FALSE))

BayesMallows documentation built on Sept. 11, 2024, 5:31 p.m.