RankDistanceModel: Fit A Mixture of Distance-based Models

In simonfqy/rankdistext:

Description

RankDistModel fits a mixture of ranking models based on weighted Kendall distance.

Usage

RankDistanceModel(dat, init, ctrl)

Arguments

dat	A RankData object.
init	A RankInit object.
ctrl	A RankControl object.

Details

The procedure will estimate central rankings, the probability of each cluster and weights.

Value

A list containing the following components:

modal_ranking.est

the estimated pi0 for each cluster.

p

the probability of each cluster.

w.est

the estimated weights of each cluster.

alpha

the estimated alpha for each cluster.

param.est

the param parametrisation of weights of each cluster.

SSR

the sum of squares of Pearson residuals

log_likelihood

the fitted log_likelihood

BIC

the fitted Bayesian Information Criterion value

free_params

the number of free parameters in the model

expectation

the expected value of each observation given by the model

iteration

the number of EM iteration

model.call

the function call

simonfqy/rankdistext documentation built on May 29, 2019, 8:19 p.m.