dfnMm {norMmix}

Number of Free Parameters of Multivariate Normal Mixture Models

Description

npar() returns an integer (vector, if p or k is) with the number of free parameters of the corresponding model, which is also the length(.) of the parameter vector in our parametrization, see nMm2par().

Usage

dfnMm(k, p, model = c("EII","VII","EEI","VEI","EVI",
	               "VVI","EEE","VEE","EVV","VVV"))

Arguments

k

number of mixture components

p

dimension of data space, i.e., number of variables (aka “features”).

model

a character string. One of the 10 models above, see also ‘Description’.

Value

integer. degrees of freedom of a model with specified dimensions, components and model type.

Examples

(m <- eval(formals(dfnMm)$model)) # list of 10 models w/ differing Sigma
# A nice table for a given 'p'  and all models, all k in 1:8
sapply(m, dfnMm, k=setNames(,1:8), p = 20)