RJi: Function to update the number of site level clusters for one...

In BAREB: A Bayesian Repulsive Biclustering Model for Periodontal Data

Description

This function updates the number of site level clusters for one patient level cluster via RJMCMC. Only used for updateRJ function.

Usage

RJi(w, K, Gamma, Beta, X, Y, Z, R, delta, mu, mu_star, c, sigma_square, C,
             theta, tau, m, n,q,T0, hyper_delta)

Arguments

w	The weights for site level clusters. Should be a vector.
K	The number of site level cluster for the patient level cluster.
Gamma	The linear coefficients for site level covariates. Should be a matrix.
Beta	The linear coefficients for patient level covariates. Should be a vector.
X	The design matrix for patient level clusters.
Y	The observed CAL value matrix
Z	The design matrix for site level clusters.
R	The current site level clustering membership
delta	The missing indicator matrix
mu	The current CAL mean vector
mu_star	The latent value vector for missingness model
c	The linear coefficients for missingness model
sigma_square	The current noise variance
C	The DPP related kernel matrices. Should be an array of 3 dimensions
theta	The DPP hyper-parameter for site level
tau	A fixed DPP hyper-parameter, which we suggest high value, say 10^5
m	The number of sites
n	The number of patients
q	The number of site level covariates
T0	The number of teeth
hyper_delta	The hyper-parameter with default value being 1

Value

RJi(w, K, Gamma, Beta, X, Y, Z, R, delta, mu, mu_star, c, sigma_square, C, theta, tau, m, n,q,T0, hyper_delta) returns a list with following variables:

K	The numbers of site level clusters
w	The weights for site level clusters
Gamma	Linear coefficients for site level covariates
R	The site level clusteirng membership
C	The DPP related kernel matrice

Author(s)

Yuliang Li

BAREB documentation built on March 26, 2020, 7:36 p.m.