aboutsummaryrefslogtreecommitdiff
AgeCommit message (Collapse)AuthorFilesLines
2022-11-18Update docsJohannes Ranke14-18/+279
2022-11-18Make mixed model test data permanent to ensure reproducibilityJohannes Ranke24-262/+324
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.
2022-11-18Round parameters with signif() before printingJohannes Ranke5-17/+24
This will hopefully make the test pass on all relevant platforms.
2022-11-18Prepare follow-up release to fix CRAN checksJohannes Ranke1-2/+2
2022-11-18Remove candidate patches that were accidentally committedJohannes Ranke2-1703/+0
2022-11-18Update NEWSJohannes Ranke1-0/+10
2022-11-18Pass error.init to saemix_model, show in parplotJohannes Ranke9-29/+51
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.
2022-11-18Avoid ill-determined fit, it is platform dependentJohannes Ranke2-12/+11
The previous test lead to an error on the Fedora test systems (gcc and clang) on CRAN: ... Results have changed from known value recorded in 'print_dfop_saemix_1.txt'. old[13:24] vs new[13:24] "" "Fitted parameters:" " estimate lower upper" - "parent_0 100.09 98.94 101.25" + "parent_0 100.17 99.01 101.33" - "log_k1 -2.68 -2.91 -2.45" + "log_k1 -2.66 -2.89 -2.43" "log_k2 -4.12 -4.24 -4.00" - "g_qlogis -0.41 -0.63 -0.20" + "g_qlogis -0.43 -0.65 -0.21" - "a.1 0.91 0.67 1.15" + "a.1 0.91 0.67 1.16" "b.1 0.05 0.04 0.06" - "SD.log_k1 0.36 0.21 0.50" + "SD.log_k1 0.35 0.21 0.50" and 2 more ...
2022-11-18Improve logLik.mkinfit to attach nobs attributeJohannes Ranke2-29/+33
The lack of that attribute lead to a failure to calculate the BIC in test_AIC.R on R-devel from yesterday.
2022-11-18Typo that did not affect test resultsJohannes Ranke1-1/+1
2022-11-17Complete update of online docs for v1.2.0Johannes Ranke166-4184/+9615
2022-11-16Reduce testing on CRAN serversv1.2.0Johannes Ranke4-6/+7
2022-11-16Update docsJohannes Ranke4-8/+11
2022-11-16Update DESCRIPTION and NEWS, check and testJohannes Ranke4-27/+27
2022-11-16Read in all data per defaultJohannes Ranke1-1/+1
2022-11-16Address platform dependence that surfaced on winbuilderJohannes Ranke2-24/+29
2022-11-16Update online docsJohannes Ranke7-127/+183
2022-11-15Import saemix for easier profiling, update online docsJohannes Ranke16-180/+76
2022-11-14Add saem benchmarks to indexJohannes Ranke7-7/+24
2022-11-14Update online docsJohannes Ranke36-1704/+1760
2022-11-14Start benchmark vignette for saem fitsJohannes Ranke4-0/+2281
2022-11-14Automatic starting parameters for saem.mmkinJohannes Ranke14-1362/+1407
For the case of mkin transformations. This gives faster convergence, and appears to avoid problems with numeric ODE solutions
2022-11-14Comment on an errorJohannes Ranke1-1/+1
I will not correct the error in order to keep the benchmarks consistent
2022-11-11Simplify travis configJohannes Ranke1-10/+1
2022-11-08Same SFORB simplification for saem.mmkin fitsJohannes Ranke1-2/+2
2022-11-08Show DFOP g equivalent of SFORB in summary, testJohannes Ranke5-29/+36
2022-11-08Simplify SFORB also in the endpoint functionJohannes Ranke3-33/+31
2022-11-08Simplify SFORB analytical solution, whitespaceJohannes Ranke4-88/+88
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.
2022-11-07Make print and summary for saem.mmkin more robustJohannes Ranke2-8/+18
If the likelihood computed by importance sampling, these methods failed. Now they report "Not available" or NA, respectively.
2022-11-04Further improvement of robustness of anova methodJohannes Ranke1-9/+9
2022-11-04Fix anova.saem.mmkin for unavailable likelihoodsJohannes Ranke1-4/+10
2022-11-04Attempt at automatic setting of random effectsJohannes Ranke7-30/+85
Based on parameters in the separate fits that fail the t-test.
2022-11-04Make it easy to override maxsteps for lsodaJohannes Ranke2-38/+44
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).
2022-11-04Make illparms.mhmkin and anova.saem.mmkin more robustJohannes Ranke3-12/+14
2022-11-04Never fail with errors in saem.mmkinJohannes Ranke3-18/+2
I think it is always preferable to be able to investigate the saemix object returned.
2022-11-03Update for cgit...Johannes Ranke1-1/+1
2022-11-03Update docsJohannes Ranke8-2/+337
2022-11-01Add tex_listing()Johannes Ranke5-1/+58
2022-11-01Add read_spreadsheetJohannes Ranke5-2/+171
2022-11-01Update online docsJohannes Ranke30-176/+275
2022-11-01Fix plotting saem fits with covariatesJohannes Ranke7-58/+94
2022-11-01Fix a bug, couldn't set class attribute on NULLJohannes Ranke1-1/+1
2022-11-01Fix illparms testsJohannes Ranke5-51/+18
2022-11-01Improved printing for illparms()Johannes Ranke7-24/+80
For the case of single fits and no ill-defined parameters found
2022-10-31Fix R CMD checkJohannes Ranke7-17/+32
2022-10-31Possibility to override the error model in update.saemJohannes Ranke2-11/+14
2022-10-31Add anova.mhmkin methodJohannes Ranke2-0/+18
2022-10-30Correct documentation of sorting in anova.saem.mmkin()Johannes Ranke1-3/+4
2022-10-30saem summary in case of non-invertible FIM partsJohannes Ranke1-7/+13
2022-10-29Another attempt to make the test robustJohannes Ranke4-20/+44
As the last update failed on travis. Now we should be more robust, as rounding to one digit works better when parameters are log-transformed (mkin transformations) in the output.

Contact - Imprint