When sadmvn, pmnorm, sadmvt, pmt
are called with dimension
larger than 20, the returned value is now NA
,
instead of a 0 with attr(,"status")="oversize"
.
The NEWS file is now in Rd format.
Function dmtruncnorm
is now based on Fortran coding of a Gibbs
sampling, removing an incorrect dependence on package tmvnsim.
Function plot_fxy
is introduced.
Now {d,p,r}mtruncnorm
can be called with parameters
which actually represent a univariate distribution.
In recintab
, biv.nt.prob
is used instead of sadmvn
when d=2
.
R version requirement is reduced from 4.0.0 to 2.2.0.
Fixed a bug which effectively prevented using dmtruncnorm
and
dmtrunct
with an argument x
of vector type.
Improvements in the documentation of recintab
and mom2cum
.
Support for the truncated versions of the multivariate normal and the t distributions is introduced.
For the truncated normal distribution, functions recintab
and
mom2cum
compute moments and cumulants.
Function mom2cum
can also be used for obtaining the
cumulants of other distributions of which the moments are known.
Evaluation of the normal and t distribution function in the trivariate case now uses specific Fortran code.
Improved R coding of sadmvn
and sadmvt
, with more
extended use of biv.nt.prob
.
Fixed a bug of pmnorm
affecting a peculiar input set.
New function sample_Mardia_measures
is introduced.
Fixed problem connected to use of 'unintialized' variable in Fortan 77 code, as claimed by valgrind.
Replaced is.matrix(.)
with code compliant with R 4.0.0
Fix a bug in rmt; this drops the earlier property that repeated calls with
a larger n
and reset set.seed(.) started with the shorter sequence.
Improved NAMESPACE and documentation.
More flexible use of mean
in {d,p,r}mnorm
and in
{d,p,r}mt
.
pd.solve
transfers dimnames(x)
to output.
rmt
and rmnorm
output with d=1
have now the same
structure.
improved checks on arguments supplied to dmnorm
and dmt
.
Introduced new argument sqrt
of dmnorm
and dmt
.
Limited check of the arguments is introduced (skipped for speed in earlier versions).
Calls to rmt
with a smaller/larger of n
generate a
sub/superset of values.
fixes a bug of parameter checking in dmnorm
and dmt
.
Improved coding of dmnorm
and dmt
functions,
avoiding use of apply
.
Allow calling pmnorm
and pmt
with a matrix argument x
.
.First.lib()
removed on request from CRAN
other minor fixes.
NAMESPACE introduced on request by the R Core Team.
fixed bug similar to the one of 1.4-2 in sadmvn
and sadmvt
.
fixed bug of biv.not.prob
if some components had
(lower=-Inf & upper=Inf)
.
Fixed inadequate documentation
coding of dmnorm
and dmt
is straightened;
log.det
in pd.solve
is returned only on request.
Function pd.solve
is introduced
argument mean
can now be a matrix for {d,p}mnorm
and {d,p}mt
.
fixed a bug affecting the case d>2 & df=Inf
.
removed check on the R version on request of the R Core team.
minor modification to Fortran 77 code to avoid GOTO jumps out of IF blocks.
evaluation of normal and t distribution functions in the bivariate case now uses specific Fortran code
improved R coding.
some amendments in documentation.
fixed a bug in function rmt
.
amedend Fortran 77 code of Alan Genz (subtroutine TRESTR).
fixed a bug of non-compliance to R programming standards.
first version uploaded on CRAN.
built private version.