Displaying 3 results from an estimated 3 matches for "dpolono".
2008 Feb 15
0
Poisson-lognormal probability calculations
...base packages:
[1] stats4 splines stats graphics grDevices utils datasets
methods base
other attached packages:
[1] VGAM_0.7-5 xlsReadWrite_1.3.2
--------------------------------
I'm having some problems with a Poisson-lognormal density (mass?) function.
VGAM has the dpolono function, but that doesn't work for x-values over 170,
and I need to go to *much* bigger numbers. It fails first because of gamma
overflow, then because of non-finite integrand.
---------------------
> VGAM::dpolono(170)
[1] 4.808781e-09
> VGAM::dpolono(171)
[1] 0
Warning message:
In VGAM...
2008 Feb 18
0
Solved (??) Behaviour of integrate (was 'Poisson-lognormal probab ility calculations')
...e
integrand "feature" mentioned in the note. I resolved it by limiting the
range of integration as shown here:
--------------------------------------------------
function (x, meanlog = 0, sdlog = 1, sdlim=6, rel.tol =
.Machine$double.eps^0.5,...) {
#+
# K. Jewell Feb 2008
# Based on VGAM::dpolono, modified to cover wider range of x
# Some simplification, but major change is to avoid integration problems by
# limiting range of integration to sdlim (default = 6) standard deviations
each side
# of mean of Poisson or lognormal. Also increase default precision of
integration.
# For x in 0:17...
2008 Feb 15
0
Behaviour of integrate (was 'Poisson-lognormal probability calcul ations')
...raphics grDevices utils
> datasets
> methods base
>
> other attached packages:
> [1] VGAM_0.7-5 xlsReadWrite_1.3.2
> --------------------------------
>
> I'm having some problems with a Poisson-lognormal density (mass?)
> function.
>
> VGAM has the dpolono function, but that doesn't work for x-values over
> 170, and I need to go to *much* bigger numbers. It fails first because
> of gamma overflow, then because of non-finite integrand.
> ---------------------
> > VGAM::dpolono(170)
> [1] 4.808781e-09
> > VGAM::dpolono(171)...