# complete.log.likelihood: Complete log-likelihood function for xCx models. In fabMix: Overfitting Bayesian Mixtures of Factor Analyzers with Parsimonious Covariance and Unknown Number of Components

## Description

Complete log-likelihood function for models with same error variance per component (xCx).

## Usage

 1 complete.log.likelihood(x_data, w, mu, Lambda, SigmaINV, z) 

## Arguments

 x_data n\times p matrix containing the data w a vector of length K containing the mixture weights mu K\times p matrix containing the marginal means per component Lambda K\times p\times q array of factor loadings per component SigmaINV p\times p precision matrix (inverse covariance) z A vector of length n containing the allocations of the n datapoints to the K mixture components

## Value

complete log-likelihood value

## Author(s)

Panagiotis Papastamoulis

## Examples

  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16  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) ) SigmaINV <- array(1, dim = c(p,p)) # compute the complete.log.likelihood complete.log.likelihood(x_data = x_data, w = w, mu = mu, Lambda = Lambda, SigmaINV = SigmaINV, z = z) 

fabMix documentation built on Feb. 20, 2020, 1:09 a.m.