Age | Commit message (Collapse) | Author | Files | Lines |
|
|
|
Thanks to Sebastian Meyer for spotting it.
|
|
|
|
|
|
|
|
|
|
Instead of rmarkdown::pdf_document, mkin::hierarchical_kinetics is used
as a document format in the template. In this way, the template file can
be freed from some R code and yaml options that the average user does
not have to be aware of.
|
|
Update docs
|
|
|
|
llquant argument, improved legend text, tests
|
|
|
|
The argument 'no_random_effect' now accepts an illparms.mhmkin object
|
|
To ensure that tests on different platforms work on the same data, the
mixed modelling test data previosly generated in
tests/testthat/setup_script.R were generated once using the script in
inst/dataset/generation/ds_mixed.R, and are now distributed with the
package.
|
|
Due to an oversight, error.init was not really passed to saemix_model in
saem.mmkin. The new initial values were reverted to c(1, 1), in order to
avoid changing the test results. Initial values for error model
parameters are now shown in parplot.multistart.
|
|
|
|
with the custom lsoda call, avoiding repeated getNativeSymbolInfo calls.
It's just that we should not be calling foreign functions from different
packages, because the may change without notice. Using
getNativeSymbolInfo for "call_lsoda" avoids the CRAN note, and a similar
call could probably be used for "unlock_solver", avoiding the NOTE in
checks for cran, but we should not do this in a CRAN package.
|
|
For the case of mkin transformations. This gives faster convergence,
and appears to avoid problems with numeric ODE solutions
|
|
|
|
|
|
I do not know why the formulae for b1 and b2 on page 64 of FOCUS
kinetics (2014) were not simplified. Clearly, the term
k12 * k21 - (k12 + k1output) * k21)
can be simplified to
- k1output * k21
The test for equivalence of DFOP and SFORB fits verifies that the change
is OK. I also removed trailing whitespaces, substituted tab characters
by two whitespaces and removed indenting of text in paragraphs
describing parameters in roxygen comments to unify formatting.
|
|
Based on parameters in the separate fits that fail the t-test.
|
|
The hope was to be more successful in saemix fits, but I did not notice
an improvement (e.g. for a failing likelihood calculation with importance
sampling, where the error message suggested to increase maxsteps).
|
|
I think it is always preferable to be able to investigate the
saemix object returned.
|
|
|
|
|
|
|
|
|
|
|
|
For the case of single fits and no ill-defined parameters found
|
|
|
|
|
|
That parhist name was not the brightest idea, as it does
not show histograms.
|
|
|
|
The reason is that it is misleading in the case of saem.mmkin objects,
because convergence is not really checked there.
|
|
|
|
|
|
The bug was introduced by the changes in summary.saem.mmkin.R and
surfaced in the tests when using saemix transformations.
|
|
Update docs
|
|
- Add 'best' and 'which.best' generics with methods for multistart
objects
- Per default, scale the parameters in parhist plots using the fit with
the highest log likelihood.
|
|
This is achieved by introducing the argument 'no_random_effect' to the
saem function.
|
|
pfm depends on mkin anyways, so reexporting set_nd_nq and
set_nd_nq_focus in pfm should provide reasonable continuity.
|
|
|
|
|
|
In particular, adapt the display of parameter boxplots
for saem fits using mkin transformations to the way
used for saem fits using saemix transformations, i.e.
always show parameters on the natural scale, and normalised them by
dividing by the median from the multiple runs.
|
|
- Split out llhist and parhist documentation
- Add example code for multistart
- Create a multistart vignette, because the example code fails when run
by pkgdown
- Fix multistart for the case of mkin transformations in the saem fit
|
|
|
|
|
|
|
|
|
|
|