observed.log.likelihood0_Sj: Log-likelihood of the mixture model

In fabMix: Overfitting Bayesian Mixtures of Factor Analyzers with Parsimonious Covariance and Unknown Number of Components

Log-likelihood of the mixture model

Description

Log-likelihood of the mixture model evaluated only at the alive components.

Usage

observed.log.likelihood0_Sj(x_data, w, mu, Lambda, Sigma, z)

Arguments

x_data	The observed data
w	Vector of mixture weights
mu	Vector of marginal means
Lambda	Factor loadings
Sigma	K\times p matrix with each row containing the diagonal of the covariance matrix of the errors per cluster
z	Allocation vector

Value

Log-likelihood value

Author(s)

Panagiotis Papastamoulis

Examples

	library('fabMix')
	data(waveDataset1500)
	x_data <- waveDataset1500[ 1:20, -1] # data
	z <-  waveDataset1500[ 1:20, 1]	# class
	p <- dim(x_data)[2]
	q <- 2
	K <- length(table(z))		# 3 classes
	# give some arbitrary values to the parameters:
	set.seed(1)
	w <- rep(1/K, K)
	mu <- array( runif(K * p), dim = c(K,p) )
	Lambda <- array( runif(K*p*q), dim = c(K,p,q) )
	Sigma <- matrix(1:K, nrow = K, ncol = p)
	# compute the complete.log.likelihood
	observed.log.likelihood0_Sj(x_data = x_data, w = w, 
		mu = mu, Lambda = Lambda, Sigma = Sigma, z = z)

fabMix documentation built on May 29, 2024, 2:53 a.m.