Skip to contents

Simulate from Multivariate (repeated measures) Mixtures of Normals

rmvnormmix.Rd

Simulate from a mixture of multivariate zero-correlation normal distributions

Usage

rmvnormmix(n, lambda=1, mu=0, sigma=1)

Arguments

n

Number of cases to simulate.

lambda

Vector of mixture probabilities with length equal to \(m\), the desired number of components. This is assumed to sum to 1; if not, it is normalized.

mu

Matrix of means of dimensions \(m\times r\), where \(m\) is the number of components (subpopulations) and \(r\) is the number of coordinates (repeated measurements) per case. Note: mu is automatically coerced to a matrix with \(m\) rows even if it is not given in this form, which can lead to unexpected behavior in some cases.

sigma

Matrix of standard deviations, same dimensions as mu. The coordinates within a case are independent, conditional on the mixture component. (There is marginal correlation among the coordinates, but this is due to the mixture structure only.) Note: sigma is automatically coerced to a matrix with \(m\) rows even if it is not given in this form, which can lead to unexpected behavior in some cases.

Details

It is possible to generate univariate standard normal random variables using the default values (but why bother?). The case of conditionally iid coordinates is covered by the situation in which all columns in mu and sigma are identical.

Value

rmvnormmix returns an \(n\times r\) matrix in which each row is a sample from one of the components of a mixture of zero-correlation multivariate normals. The mixture structure induces nonzero correlations among the coordinates.

See also

rnormmix

Examples

##Generate data from a 2-component mixture of trivariate normals.

set.seed(100)
n <- 200
lambda <- rep(1, 2)/2
mu <- matrix(2*(1:6), 2, 3)
sigma <- matrix(1,2,3)
mydata<-rmvnormmix(n, lambda, mu, sigma)

## Now check to see if we can estimate mixture densities well:
title <- paste("Does this resemble N(", mu[1,], ",1) and N(", mu[2,],",1)?",
                sep="")
plot(npEM(mydata, 2), title=title)
#> iteration 1   lambda  0.515 0.485 time 0.001 
#> iteration 2   lambda  0.5078 0.4922 time 0.001 
#> iteration 3   lambda  0.5018 0.4982 time 0.001 
#> iteration 4   lambda  0.4969 0.5031 time 0.001 
#> iteration 5   lambda  0.4932 0.5068 time 0.002 
#> iteration 6   lambda  0.4903 0.5097 time 0.001 
#> iteration 7   lambda  0.4881 0.5119 time 0.001 
#> iteration 8   lambda  0.4864 0.5136 time 0.001 
#> iteration 9   lambda  0.4851 0.5149 time 0.002 
#> iteration 10   lambda  0.4842 0.5158 time 0.001 
#> iteration 11   lambda  0.4834 0.5166 time 0.001 
#> iteration 12   lambda  0.4829 0.5171 time 0.002 
#> iteration 13   lambda  0.4824 0.5176 time 0.001 
#> iteration 14   lambda  0.4821 0.5179 time 0.001 
#> iteration 15   lambda  0.4818 0.5182 time 0.002 
#> iteration 16   lambda  0.4816 0.5184 time 0.001 
#> iteration 17   lambda  0.4815 0.5185 time 0.001 
#> iteration 18   lambda  0.4814 0.5186 time 0.002 
#> iteration 19   lambda  0.4813 0.5187 time 0.001 
#> iteration 20   lambda  0.4812 0.5188 time 0.001 
#> iteration 21   lambda  0.4812 0.5188 time 0.001 
#> iteration 22   lambda  0.4811 0.5189 time 0.002 
#> iteration 23   lambda  0.4811 0.5189 time 0.001 
#> iteration 24   lambda  0.4811 0.5189 time 0.001 
#> iteration 25   lambda  0.4811 0.5189 time 0.001 
#> iteration 26   lambda  0.4811 0.5189 time 0.002 
#> iteration 27   lambda  0.4811 0.5189 time 0.001 
#> iteration 28   lambda  0.481 0.519 time 0.001 
#> iteration 29   lambda  0.481 0.519 time 0.002 
#> iteration 30   lambda  0.481 0.519 time 0.001 
#> iteration 31   lambda  0.481 0.519 time 0.001 
#> iteration 32   lambda  0.481 0.519 time 0.002 
#> iteration 33   lambda  0.481 0.519 time 0.001 
#> iteration 34   lambda  0.481 0.519 time 0.001 
#> iteration 35   lambda  0.481 0.519 time 0.002 
#> iteration 36   lambda  0.481 0.519 time 0.003 
#> iteration 37   lambda  0.481 0.519 time 0.001 
#> iteration 38   lambda  0.481 0.519 time 0.002 
#> iteration 39   lambda  0.481 0.519 time 0.001 
#> iteration 40   lambda  0.481 0.519 time 0.001 
#> iteration 41   lambda  0.481 0.519 time 0.001 
#> iteration 42   lambda  0.481 0.519 time 0.002 
#> iteration 43   lambda  0.481 0.519 time 0.001 
#> iteration 44   lambda  0.481 0.519 time 0.001 
#> iteration 45   lambda  0.481 0.519 time 0.001 
#> iteration 46   lambda  0.481 0.519 time 0.002 
#> iteration 47   lambda  0.481 0.519 time 0.001 
#> iteration 48   lambda  0.481 0.519 time 0.001 
#> iteration 49   lambda  0.481 0.519 time 0.002 
#> iteration 50   lambda  0.481 0.519 time 0.001 
#> iteration 51   lambda  0.481 0.519 time 0.001 
#> iteration 52   lambda  0.481 0.519 time 0.002 
#> iteration 53   lambda  0.481 0.519 time 0.001 
#> lambda  0.481 0.519, total time 0.071 s