gbdcd_regression: Gaussian BDCD regression

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

Description

Implementation of the Bayesian Detection of Clusters and Discontinuities

Usage

gbdcd_regression(
  y,
  x,
  viz,
  n_iterations = 1e+05,
  burn_in = 50000,
  c = 0.35,
  prior_coeffs_mu = rep(0, 2),
  prior_sigma = sqrt(2),
  lambda = 0.001,
  plot = F
)

Arguments

y	a vector specifying the target variable for each of the nodes.
x	a matrix with the independent variables.
viz	a matrix defining the graph neighbors.
n_iterations	number of iterations.
burn_in	number of discarded iterations.
c	parameter indicating the a priori number of clusters.
prior_coeffs_mu	a priori mean for regresion coefficients.
prior_sigma	a priori standard deviation.
lambda	regularization parameter.
plot	plot the results. Default = FALSE.

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()

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