run_gibbs: Runs the Gibbs sampler algorithm using using initial values...

In BayesCPclust: A Bayesian Approach for Clustering Constant-Wise Change-Point Data

Runs the Gibbs sampler algorithm using using initial values for the parameters

Description

Runs the Gibbs sampler algorithm using using initial values for the parameters

Usage

run_gibbs(
  M,
  N,
  w,
  d,
  as = 2,
  bs = 100,
  al = 2,
  bl = 1000,
  a = 2,
  b = 1000,
  alpha0 = 1/100,
  lambda = 2,
  maxIter = 10000,
  par.values,
  data,
  cluster,
  sigma2
)

Arguments

M	A scalar representing the number of points available for each observation
N	A scalar representing the number of observations
w	A scalar representing the minimum number of points in each interval between two change points
d	A scalar representing the number of clusters.
as	The hyperparameter value for the shape parameter in the inverse-gamma prior for the variance component
bs	The hyperparameter value for the scale parameter in the inverse-gamma prior for the variance component
al	The hyperparameter value for the shape parameter in the gamma prior for lambda
bl	The hyperparameter value for the scale parameter in the gamma prior for lambda
a	The hyperparameter value for the shape parameter in the gamma prior for alpha0
b	The hyperparameter value for the scale parameter in the gamma prior for alpha0
alpha0	A scalar defining the parameter for the Dirichlet process prior that controls the number of clusters (or its initial values)
lambda	A scalar defining the parameter for the Truncate Poisson distribution that controls the number of change points (or its initial values)
maxIter	A scalar for the number of iteration to run in the Gibbs sampler
par.values	A list with lists with parameters for each cluster. The first argument in each list is the number of change points, then the positions for the change points, where T_1 = 1, T_last = M + 1, and for each interval between change points you need to specify a value for the constant level. If running the Gibbs sampler for a dataset with unknown number of change points, we suggest setting the number of change points for each cluster to be zero. Check example in README file.
data	a matrix of size M x N with data sequences in the columns
cluster	a vector with cluster assignments for each data sequence
sigma2	a vector with variance components for each data sequence

Value

A list with estimates for each iteration of the Gibbs sampler for each parameter

Examples

d = 2 # two clusters
N = 5 # 5 data sequences
M = 50 # 50 observations for each data sequence
maxIter = 10 # number of Gibbs sampler iterations

data(data)
# initial values for each paramter and each cluster
par.values <- list(K = c(0, 0), Tl = list(50, 50), alpha = list(5, 10))
#cluster assignment for each data sequence
cluster <- kmeans(t(data), 2)$cluster
# variance for each data sequence
sigma2 <- apply(data, 2, var)
res <- run_gibbs(M, N, w = 10, d, as = 2, bs = 100, al = 2, bl = 1000, a = 2,
 b = 1000, alpha0 = 1/100, lambda = 2, maxIter = 10, par.values, data,
 cluster, sigma2)

BayesCPclust documentation built on April 4, 2025, 5:19 a.m.