v2.1.0
nlmixr2est 2.1.0
Breaking changes
FOCEi
-
Gill forward differences will not repeat now (by default), You can
change back to prior behavior withfoceiControl(repeatGillMax=3)
-
Number of sticky recalculation is reduced to 4; to have the old
behavior usefoceiControl(stickyRecalcN=5)
-
n2ll
has been changed toll
to specify individual
log-likelihoods. This was only used in simulation and was not well
documented. -
Generalized log-likelihood is only supported with
rxode2
2.0.8
or later.
FOCEi covariance calculation
-
The
S
matrix calculation was made a bit more robust to errors in
individual gradients. When there are errors in the individual
gradient calculation, assume the gradient is the same as the
overall gradient. In the tests cases, were reasonable using this
adjusted S matrix. This means if some individuals do not have very
much data to support a specific parameter, aS
matrix calculation
for the population will still be generated. When there is some
patients/subject combinations that do not have sufficient data, we
will add the following to the run information:S matrix had problems solving for some subject and parameters
. TheS
matrix
calculation will still fail if the percentage of parameters that
are being reset is lower thanfoceiControl(smatPer=0.6)
or
whatever you specify. -
The
r,s
covariance matrix will now also check for unreasonably
small values (controlled byfoceiControl(covSmall=...)
) and
select a different covariance estimate method even when the "r" and
"s" matrices are calculated "correctly".
New features
-
What type(s) censoring (if any) is now stored in
fit$censInformation
-
Standard errors of
$etas
can now be obtained withfit$phiSE
,
also available arefit$phiRSE
(relative standard error),
fit$phiH
, (individual hessian),fit$phiC
(individual
covariances),fit$phiR
(individual correlation matrices) -
Can also use Shi 2021 differences in addition to Gill differences.
In our tests (using the same datasets as CPT) these produced worse
estimates than the Gill 1983, though it is unclear why since it
should be a faster more accurate method. A modified version is used
in calculating the individual Hessians of numerically for the
generalized likelihood approach. -
Generalized likelihood estimation is now present in
nlmixr2est
for
focei
,foce
andposthoc
-
nmNearPD()
is a function you may use for nearest positive definite
matrix. This is derived fromMatrix::nearPD()
but is implemented
in C/C++ to be used in (possibly threaded) optimization. -
Individual Hessians can be accessed by
$phiH
, covariance by
$phiC
, eta standard errors by$phiSE
and eta RSEs can be
accessed by$phiRSE
. There areeta
aliases for these as well
($etaH
,$etaC
,$etaSE
, and$etaRSE
). -
Can now access the individual point's contribution to the overall
likelihood when merging to the original dataset. These merges can be
accessed with$dataMergeFull
,$dataMergeLeft
,$dataMergeRight
,
and$dataMergeInner
. The columns with the individual data column
isnlmixrLlikObs
.To calculate the total
focei
/foce
objective function, the sum of the
likelihoods still need to be adjusted by the omega/eta contribution,
and the individual Hessians, and possibly the NONMEM objective
function offset constant.
Censoring fixes
- Fixed bug where datasets with censoring that are not lower case
cens
andlimit
do not
produce the correct table output (#180)
FOCEi updates
- Resets now scale properly when a value is simulated outside the limit
- Models with zero gradients on the first step now switch to
bobyqa
by default. With this, it is more important to examine the model
parameters and fits for plausibility.
What's Changed
- 179 m4 method simulations may be off by @mattfidler in #182
- 151 when i just tried to simulate from my model and data i got the below errors im digging deeper to try to figure out what is the matter by @mattfidler in #187
- Add a different stop so that it gives a bit more information when it … by @mattfidler in #191
- Llik focei by @mattfidler in #188
- 184 feature request move $warnings to $runinfo or something similar and dont warn by @mattfidler in #193
- 200 rcpparmadillo 2 by @mattfidler in #203
- 205 function rxgt is not supported in rxode2 by @mattfidler in #206
- Print out every time by @mattfidler in #207
- Add $dataLloq and $dataUloq to fit by @mattfidler in #212
- Keep vpcSim the same if it is normally related items by @mattfidler in #209
- Fix simulation to support dvid() and models without dvid()/cmt() by @mattfidler in #214
- Fix for vpc sim by @mattfidler in #216
- Fall back to length normalization type by @mattfidler in #219
- 217 rse focei less stable than last release by @mattfidler in #220
- Make sure the etas are fixed with the output by @mattfidler in #224
- 199 running a model changes the source object by @mattfidler in #204
- 226 restart error by @mattfidler in #227
- 228 preds seem wrong when mixing normal log likelihood and non normal log likelihood by @mattfidler in #229
- Now test pred, ipred; pred fails by @mattfidler in #230
- Start splitting out depending on what is available with rxode2 by @mattfidler in #232
- Address issues for nlmixr2est by @mattfidler in #234
- Add new CRAN submission tag by @mattfidler in #235
- Fix split model by @mattfidler in #237
- Add some split testing by @mattfidler in #239
- Check for environment a bit more stricter by @mattfidler in #242
- Handle edge cases for NA predictions by @mattfidler in #243
- Env is compressed by @mattfidler in #245
- 246 nlmixr2est with rxode2parse by @mattfidler in #247
- 246 nlmixr2est with rxode2parse by @mattfidler in #248
- 250 rxode2random by @mattfidler in #251
- 252 use rxode2et by @mattfidler in #253
- Fix and test for non mu-referenced model by @mattfidler in #259
- Make table integration more robust by @billdenney in #256
- Make .foceiPreProcessData() work with tibbles by @billdenney in #262
Full Changelog: v2.0.8...v2.1.0