mkin/man, branch v0.9.49.5Fitting kinetic models to chemical degradation data (also on github)https://cgit.jrwb.de/mkin/atom?h=v0.9.49.52019-07-02T15:35:49ZTypo2019-07-02T15:35:49ZJohannes Rankejranke@uni-bremen.de2019-07-02T15:35:49Zurn:sha1:e09b0a1ac18cc5e5c90b06539868ebde56131db8Add sources for test data for better transparency2019-06-04T17:12:47ZJohannes Rankejranke@uni-bremen.de2019-06-04T17:12:47Zurn:sha1:e36a5db313365bc5acc85126792c767fe18acd71Additional algorithm "d_c", more tests, docs2019-06-04T13:09:28ZJohannes Rankejranke@uni-bremen.de2019-06-04T13:09:28Zurn:sha1:95178837d3f91e84837628446b5fd468179af2b9
The new algorithm tries direct optimization of the likelihood, as well
as a three step procedure. In this way, we consistently get the
model with the highest likelihood for SFO, DFOP and HS for all 12
new test datasets.
Experimental data for finding error model algorithm2019-05-31T12:32:19ZJohannes Rankejranke@uni-bremen.de2019-05-31T12:32:19Zurn:sha1:037cdd16f39b8b889e7bda26961a90cd68c6f5a9Clarify the relation of the two-component error model2019-05-21T06:54:16ZJohannes Rankejranke@uni-bremen.de2019-05-21T06:54:16Zurn:sha1:84abde72967691d82bbad66eeff4d1ab161530dd
in relation to the original version by Rocke and Lorenzato
Add functionality to plot the error model2019-05-08T18:57:48ZJohannes Rankejranke@uni-bremen.de2019-05-08T18:57:48Zurn:sha1:c6079a807e2b400fe0c772603392aeacd887da2f
by plotting squared residuals against predicted values, and
showing the variance function used in the fitted error model.
Rebuild docs
Prepare for CRAN release2019-05-02T15:51:07ZJohannes Rankejranke@uni-bremen.de2019-05-02T15:07:55Zurn:sha1:a4ca3451f1b5c37d10c6a41cb18a99b1631e8aa2
- Skip long running tests on CRAN as well to avoid timeout on winbuilder
- Don't install benchmark results in the package, they are only needed
in the git repository
- Don't run example in man/add_err.Rd as it takes > 10 s on winbuilder
- Rebuild docs
Better initials for error model parameters2019-05-02T11:17:05ZJohannes Rankejranke@uni-bremen.de2019-05-02T11:17:05Zurn:sha1:70591022c07f0e8fb4dd67789b7c8d78af8ebc18
- Also make it possible to specify initial values for error model
parameters.
- Run tests
- Rebuild docs
Adapt tests, vignettes and examples2019-04-10T08:17:35ZJohannes Rankejranke@uni-bremen.de2019-04-10T08:17:35Zurn:sha1:194659fcaccdd1ee37851725b8c72e99daa3a8cf
- Write the NEWS
- Static documentation rebuilt by pkgdown
- Adapt mkinerrmin
- Fix (hopefully all) remaining problems in mkinfit
Direct error model fitting works2019-04-04T15:21:13ZJohannes Rankejranke@uni-bremen.de2019-04-04T13:42:23Zurn:sha1:7a1d3d031aa23fce723ac4f4c8e4bb5d64959447
- No IRLS required
- Removed optimization algorithms other than Port
- Removed the dependency on FME
- Fitting the error model 'obs' is much faster for the FOCUS_2006_D
dataset and the FOMC_SFO model (1 second versus 3.4 seconds)
- Vignettes build slower. Compiled models needs 3 minutes instead of 1.5
- For other vignettes, the trend is less clear. Some fits are faster,
even for error_model = "const". FOCUS_Z is faster (34.9 s versus
44.1 s)
- Standard errors and confidence intervals are slightly smaller
- Removed code for plotting during the fit, as I hardly ever used it
- Merged the two cost functions (using transformed and untransformed
parameters) into one log-likelihood function