correction_kernel_pseudo: Correction Kernel (pseudolikelihood)

In BayesMallows: Bayesian Preference Learning with the Mallows Rank Model

Correction Kernel (pseudolikelihood)

Description

Function to determine if the augmented ranking is compatible with the new observed partial ranking. If it is not, the we create a new augmentation using the pseudolikelihood approach and calculate the augmentation probability.

Usage

correction_kernel_pseudo(
  current_ranking,
  observed_ranking,
  rho,
  alpha,
  n_items,
  metric = "footrule"
)

Arguments

current_ranking	An complete rank sequence vector of the proposed augmented ranking obtained from calculate_forward_probability function
observed_ranking	An incomplete rank sequence vector of the original observed incomplete ranking which contains NAs
rho	Numeric vector specifying the consensus ranking
alpha	Numeric value of the scale parameter
n_items	Integer is the number of items in a ranking
metric	A character string specifying the distance metric to use in the Bayesian Mallows Model. Available options are "footrule", "spearman", "cayley", "hamming", "kendall", and "ulam".

Value

list containing R_obs, the proposed 'corrected' augmented ranking that is compatible with the new observed ranking for a user, and forward_auxiliary_ranking_probability, a numerical value for the probability of correcting the ranking to be compatible with R_obs.

BayesMallows documentation built on Nov. 25, 2023, 5:09 p.m.