Re: [R] Issues with 'Miwa' algorithm in mvtnorm package

2017-10-02 Thread Hollie Johnson (PGR)
:13 To: Bert Gunter Cc: r-help@r-project.org; Hollie Johnson (PGR) Subject: Re: [R] Issues with 'Miwa' algorithm in mvtnorm package Hi Hollie, I tried to reproduce your example but could not. Here is what I did. Please explain if you did something different. x <- c(9.358*10^-3, -8.165*

Re: [R] Issues with 'Miwa' algorithm in mvtnorm package

2017-10-02 Thread Eric Berger
; >> ericjber...@gmail.com> >> *Sent:* 02 October 2017 16:16:13 >> *To:* Bert Gunter >> *Cc:* r-help@r-project.org; Hollie Johnson (PGR) >> *Subject:* Re: [R] Issues with 'Miwa' algorithm in mvtnorm package >> >> Hi Hollie, >> I

Re: [R] Issues with 'Miwa' algorithm in mvtnorm package

2017-10-02 Thread Eric Berger
) should read B <- matrix(y, nrow=3, > byrow = TRUE) > > > Regards, Hollie > -- > *From:* R-help on behalf of Eric Berger < > ericjber...@gmail.com> > *Sent:* 02 October 2017 16:16:13 > *To:* Bert Gunter > *Cc:* r-help@r-projec

Re: [R] Issues with 'Miwa' algorithm in mvtnorm package

2017-10-02 Thread Eric Berger
Hi Hollie, I tried to reproduce your example but could not. Here is what I did. Please explain if you did something different. x <- c(9.358*10^-3, -8.165*10^-3, -1.689*10^-8, -8.165*10^-3, 9.358*10^-3, 1.854*10^-8, -1.689*10^-8, 1.854*10^-8, 9.358*10^-3) A <- matrix(x, nrow=3

Re: [R] Issues with 'Miwa' algorithm in mvtnorm package

2017-10-02 Thread Bert Gunter
Rather specialized. As this appears to be primarily a statistical, not an R programming question, you may do better posting on a statistical site like http://stats.stackexchange.com/ if you don't get a satisfactory reply here . Alternativey, if you think this is a package bug, perhaps contact the

[R] Issues with 'Miwa' algorithm in mvtnorm package

2017-10-02 Thread Hollie Johnson (PGR)
Currently doing some work on local maxima on a random field and have encountered an issue with the Miwa algorithm used with the pmvnorm function in the mvtnorm R package. Based on recommendations by Mi et al., we ran the mvtnorm package using the Miwa algorithm, since we have a maximum of 4 dim