mbc: Fit a mixture of multivariate Gaussians

Description Usage Arguments Value Examples

View source: R/mbc.R

Description

Fit a mixture of multivariate Gaussians

Usage

1
2
3
4
5
6
mbc(x, groups = 2, maxiter = 500, likelihood = TRUE, verbose = FALSE,
  plot = FALSE, z = NULL)

mbc_cond(x_A, x_B, mean_A, sigma_AA, z, pro, groups, maxiter = 500,
  likelihood = TRUE, verbose = FALSE, plot = FALSE, abstol = 0.001,
  method_sigma_AB = c("analytic", "numeric"), updateA = FALSE)

Arguments

x, x_A, x_B

Data frame or a matrix

groups

The number of groups/mixture components to fit.

maxiter

The maximum number of iterations for the E-M algorithm.

likelihood

Logical indicating whether the log-likelihood should be calculated at each step and returned (defaults to TRUE).

verbose

Print verbose output if TRUE.

plot

Visualise the mixture model as it progresses.

z

A matrix of cluster probabilities.

mean_A

Mean vectors for previous batch.

sigma_AA

Covariance matrices for previous batch.

pro

Mixing proportions.

abstol

Stopping tolerance for likelihood.

method_sigma_AB

One of "analytic" (default) or "numeric", to choose the method of estimating the between-batch covariance for a cluster.

updateA

Logical, if TRUE, updates the parameters for the previous batch after estimating the paramters for the current batch.

Value

A list containing the estimated parameters for the mixture distribution.

Examples

1
2
3
library(mclust)
data(banknote)
mbc(x = banknote[, -1], groups = 2)

markajoc/MBCbigP documentation built on July 12, 2017, 12:23 a.m.