R help - Feb 2011 - Question about checkTmvArgs function in rtmvnorm (package tmvtnorm)

Hello!

I was wondering if it's possible to see the actual code of
checkTmvArgs function that is part of the code for rtmvnorm (which is
below - I just typed "rtmvnorm" on the prompt). I get an error:

Error in checkTmvArgs(mean, sigma, lower, upper) :
  sigma must be a symmetric matrix

At the same time I am pretty sure that the matrix I am passing as
sigma is a var-covar matrix (however, it is NOT a correlation matrix).

Thanks a lot!
Dimitri


function (n, mean = rep(0, nrow(sigma)), sigma = diag(length(mean)),
    lower = rep(-Inf, length = length(mean)), upper = rep(Inf,
        length = length(mean)), algorithm = c("rejection",
"gibbs",
        "gibbsR"), ...)
{
    algorithm <- match.arg(algorithm)
    cargs <- checkTmvArgs(mean, sigma, lower, upper)
    mean <- cargs$mean
    sigma <- cargs$sigma
    lower <- cargs$lower
    upper <- cargs$upper
    if (n < 1 || !is.numeric(n) || n != as.integer(n) || length(n) >
        1) {
        stop("n must be a integer scalar > 0")
    }
    if (algorithm == "rejection") {
        retval <- rtmvnorm.rejection(n, mean, sigma, lower, upper,
            ...)
    }
    else if (algorithm == "gibbs") {
        retval <- rtmvnorm.gibbs.Fortran(n, mean, sigma, lower,
            upper, ...)
    }
    else if (algorithm == "gibbsR") {
        retval <- rtmvnorm.gibbs(n, mean, sigma, lower, upper,
            ...)
    }
    return(retval)
}

-- 
Dimitri Liakhovitski
Ninah Consulting
www.ninah.com

I found a reason why I was getting the error that my var-covar matrix
was not symmetric: because my column names and row names were
different!
Dimitri

On Mon, Feb 7, 2011 at 12:39 PM, Dimitri Liakhovitski
<dimitri.liakhovitski at gmail.com> wrote:
> Hello!
>
> I was wondering if it's possible to see the actual code of
> checkTmvArgs function that is part of the code for rtmvnorm (which is
> below - I just typed "rtmvnorm" on the prompt). I get an error:
>
> Error in checkTmvArgs(mean, sigma, lower, upper) :
> ?sigma must be a symmetric matrix
>
> At the same time I am pretty sure that the matrix I am passing as
> sigma is a var-covar matrix (however, it is NOT a correlation matrix).
>
> Thanks a lot!
> Dimitri
>
>
> function (n, mean = rep(0, nrow(sigma)), sigma = diag(length(mean)),
> ? ?lower = rep(-Inf, length = length(mean)), upper = rep(Inf,
> ? ? ? ?length = length(mean)), algorithm = c("rejection",
"gibbs",
> ? ? ? ?"gibbsR"), ...)
> {
> ? ?algorithm <- match.arg(algorithm)
> ? ?cargs <- checkTmvArgs(mean, sigma, lower, upper)
> ? ?mean <- cargs$mean
> ? ?sigma <- cargs$sigma
> ? ?lower <- cargs$lower
> ? ?upper <- cargs$upper
> ? ?if (n < 1 || !is.numeric(n) || n != as.integer(n) || length(n) >
> ? ? ? ?1) {
> ? ? ? ?stop("n must be a integer scalar > 0")
> ? ?}
> ? ?if (algorithm == "rejection") {
> ? ? ? ?retval <- rtmvnorm.rejection(n, mean, sigma, lower, upper,
> ? ? ? ? ? ?...)
> ? ?}
> ? ?else if (algorithm == "gibbs") {
> ? ? ? ?retval <- rtmvnorm.gibbs.Fortran(n, mean, sigma, lower,
> ? ? ? ? ? ?upper, ...)
> ? ?}
> ? ?else if (algorithm == "gibbsR") {
> ? ? ? ?retval <- rtmvnorm.gibbs(n, mean, sigma, lower, upper,
> ? ? ? ? ? ?...)
> ? ?}
> ? ?return(retval)
> }
>
> --
> Dimitri Liakhovitski
> Ninah Consulting
> www.ninah.com
>


-- 
Dimitri Liakhovitski
Ninah Consulting
www.ninah.com