qincolmult: Convert multiple symmetric matrices into vectors

In simonsays1980/finmix: An R package for Bayesian estimation of finite mixture distributions

Description

Calling qincolmult() on an array of symmetric matrices all with dimension rxr converts these matrices into an array of vectors with length r(r+1)/2. This function is used to handle the MCMC sampling output from multivariate finite mixture models. To save storage the symmetric variance-covariance matrices of multivariate mixtures are stored in vector form. If the covariance matrices are needed for calculations the functions qinmatr() and qinmatrmult() helps to restore these matrices from the storage vectors.

Usage

qincolmult(a)

Arguments

a	An array of symmetric matrices or dimension rxrxK.

Value

A matrix of dimension r(r+1)/2xK.

See Also

• qinmatr() for converting a single vector into a symmetric matrix

• qinmatrmult() for converting multiple vectors into symmetric matrices

• qincol() for converting a single symmetric matrix into a vector

Examples

# Convert a matrix of vectors
matrices <- qinmatrmult(matrix(rnorm(36), nrow = 6))
# Convert these matrices back into vectors.
qincolmult(matrices) 

simonsays1980/finmix documentation built on Dec. 23, 2021, 2:25 a.m.