Skip to contents

Specify various model options for the Bayesian Mallows model.


  metric = c("footrule", "spearman", "cayley", "hamming", "kendall", "ulam"),
  n_clusters = 1,
  error_model = c("none", "bernoulli")



A character string specifying the distance metric to use in the Bayesian Mallows Model. Available options are "footrule", "spearman", "cayley", "hamming", "kendall", and "ulam". The distance given by metric is also used to compute within-cluster distances, when include_wcd = TRUE.


Integer specifying the number of clusters, i.e., the number of mixture components to use. Defaults to 1L, which means no clustering is performed. See compute_mallows_mixtures() for a convenience function for computing several models with varying numbers of mixtures.


Character string specifying which model to use for inconsistent rankings. Defaults to "none", which means that inconsistent rankings are not allowed. At the moment, the only available other option is "bernoulli", which means that the Bernoulli error model is used. See Crispino et al. (2019) for a definition of the Bernoulli model.


An object of class "BayesMallowsModelOptions", to be provided in the model_options argument to compute_mallows(), compute_mallows_mixtures(), or update_mallows().


Crispino M, Arjas E, Vitelli V, Barrett N, Frigessi A (2019). “A Bayesian Mallows approach to nontransitive pair comparison data: How human are sounds?” The Annals of Applied Statistics, 13(1), 492--519. doi:10.1214/18-aoas1203 .