gbdcd: Gaussian BDCD

In leandromineti/gbdcd: Implementation of the Bayesian Detection of Clusters and Discontinuities

Description

Implementation of the Bayesian Detection of Clusters and Discontinuities

Usage

gbdcd(
  y,
  neigh,
  c = 0.35,
  n_iterations = 1e+06,
  burn_in = 5e+05,
  mu0 = 0,
  sigma_0 = sqrt(2)
)

Arguments

y	a vector specifying the target variable for each of the nodes.
neigh	a matrix defining the graph neighbors.
c	parameter indicating the a priori number of clusters.
n_iterations	number of iterations.
burn_in	number of discarded iterations.
mu0	a priori mean.
sigma_0	a priori standard deviation.

Value

a list of seven objects:

• mean.info: a posteriori means and credible interval.

• cluster.info: hierarchical clustering return object.

• matConnections: frequency matrix indicating how many times each pair of nodes was in the same cluster.

• k.MCMC: a vector indicating the number of clusters in each iteration.

• mean_y: target variable mean.

• sd_y: target variable standard deviation.

• vec.centers: center configuration for each iteration.

Author(s)

[e-mail]

See Also

Other gbdcd: gbdcdGroups(), gbdcd_regression()

Examples

library(gbdcd)

data("aneeldata", package = "gbdcd")
data("aneelshape", package = "gbdcd")

target_variable <- aneelshape$z_Precipitation
neighbors <- aneeldata$connections

out <- gbdcd(
  y = target_variable,
  neigh = neighbors,
  c = 0.35,
  n_iterations = 100000,
  burn_in = 50000,
  mu0 = 0,
  sigma_0 = sqrt(2)
)

leandromineti/gbdcd documentation built on Sept. 24, 2020, 12:43 p.m.