rtnorm
now checks arguments length and throw an error rather than try to recycle mean and standard deviation vectors (issue #6).rtnorm
recycles arguments when ltrunc
or rtrunc
are length 1.*tmvt
and *tmvnorm
now check that the scale matrix sigma
is symmetric, positive-definite and non-degenerate (issue #8) by default.rmvnorm
returns a vector rather than an 1 by n matrix for the unidimensional setting.tinytest
rather than testthat
for unit tests.randtoolbox
(throwing warning messages), the package now relies on the qrng
package and the Owen scrambling method from spacefillr
package for quasi Monte Carlo.log
(never implemented) from functions pmvnorm
and pmvt
; these now return a warning (thanks to Elysée Aristide)ptmvnorm
and ptmvt
now return 1 if all points exceed the upper bounds (incorrectly returned zero in v2.1.2 and NA
in previous versions) (thanks to Adelchi Azzalini).ptmvtnorm
and ptmvt
return 0 for points outside the truncated region instead of NA
(thanks to Adelchi Azzalini)rtmvt
that returned incorrect output (tranposed mean vector)lb
less than ub
, to avoid segfault errors. In case of degenerate bounds, the conditional distribution is used with fixed components for the degenerate variables (issue #2)randtoolbox
package following its reuploading on CRANNaN
in boundsmvrandt
and mvrandn
randtoolbox
package unmaintained), but the sequence is not scramblednleq
solver to solve convex program (also check KKT conditions).Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.