Age | Commit message (Collapse) | Author | Files | Lines | |
---|---|---|---|---|---|
2019-09-18 | Correct and rebuild docs | Johannes Ranke | 1 | -9/+21 | |
- Reconcile docs and code for max_twa_parent - Correct links to docs in twa vignette - Static documentation rebuilt by pkgdown | |||||
2019-09-02 | Some changes to improve plots on beamer slides | Johannes Ranke | 3 | -15/+22 | |
2019-07-09 | Adaptations for gmkin | Johannes Ranke | 1 | -1/+3 | |
Address winbuilder check problems, update check log, update of static docs | |||||
2019-07-08 | Support summarizing older fits | Johannes Ranke | 1 | -8/+10 | |
generated with mkin < 0.9.49.5 | |||||
2019-07-05 | Static documentation rebuilt by pkgdown | Johannes Ranke | 1 | -0/+1 | |
2019-06-05 | Add error model algorithm to output | Johannes Ranke | 1 | -8/+19 | |
2019-06-04 | Fix a bug introduced in the last commit | Johannes Ranke | 1 | -2/+4 | |
2019-06-04 | For the d_3 algorithm, report which was better, if any | Johannes Ranke | 1 | -2/+13 | |
2019-06-04 | Additional algorithm "d_c", more tests, docs | Johannes Ranke | 1 | -46/+71 | |
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. | |||||
2019-06-04 | Algorithms direct, two-, three-, fourstep, IRLS | Johannes Ranke | 1 | -9/+41 | |
All of them are working now and allow for comparison Based on SFO, DFOP and HS fits to twelve test datasets, only the combination of direct and threestep is needed to find the lowest AIC | |||||
2019-06-03 | Status von Samstag morgen - untested | Johannes Ranke | 1 | -37/+63 | |
2019-05-23 | Fix a bug that could lead to wrong chi2 error levels | Johannes Ranke | 1 | -0/+1 | |
2019-05-20 | Only use colors for error model "obs" | Johannes Ranke | 1 | -3/+3 | |
2019-05-08 | Add functionality to plot the error model | Johannes Ranke | 4 | -15/+116 | |
by plotting squared residuals against predicted values, and showing the variance function used in the fitted error model. Rebuild docs | |||||
2019-05-07 | Fix hessian calculations | Johannes Ranke | 1 | -7/+8 | |
Static documentation rebuilt by pkgdown | |||||
2019-05-07 | After the OLS step, use OLS parameter estimates | Johannes Ranke | 1 | -24/+24 | |
- Fix the respective error in the code - Static documentation rebuilt by pkgdown | |||||
2019-05-03 | Improve error model fitting | Johannes Ranke | 2 | -15/+36 | |
Now we have a three stage fitting process for nonconstant error models: - Unweighted least squares - Only optimize the error model - Optimize both Static documentation rebuilt by pkgdown | |||||
2019-05-02 | Better initials for error model parameters | Johannes Ranke | 1 | -8/+21 | |
- Also make it possible to specify initial values for error model parameters. - Run tests - Rebuild docs | |||||
2019-04-24 | Improve output and update tests | Johannes Ranke | 1 | -5/+9 | |
Remove skipped tests as I do not intend to reactivate them | |||||
2019-04-10 | Address failure on travis | Johannes Ranke | 1 | -1/+1 | |
Also reduce the digits in the representative half-live given by nafta() | |||||
2019-04-10 | Adapt tests, vignettes and examples | Johannes Ranke | 3 | -68/+90 | |
- Write the NEWS - Static documentation rebuilt by pkgdown - Adapt mkinerrmin - Fix (hopefully all) remaining problems in mkinfit | |||||
2019-04-08 | Remove zero observations, adapt logLik | Johannes Ranke | 2 | -24/+11 | |
Zero observations at time zero made fitting the two-component error model fail. A concentration of exactly zero does not make sense anyways, as we generally have a limit of detection | |||||
2019-04-05 | Report negative log-likelihood during fits | Johannes Ranke | 1 | -4/+10 | |
If not quiet = TRUE | |||||
2019-04-04 | Direct error model fitting works | Johannes Ranke | 1 | -366/+133 | |
- 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 | |||||
2019-03-04 | signif for DTx when printing nafta results | Johannes Ranke | 1 | -1/+1 | |
In order to be able to test cross-platform | |||||
2019-03-04 | Review tests and examples based on winbuilder results | Johannes Ranke | 1 | -3/+3 | |
Static documentation rebuilt by pkgdown | |||||
2019-02-27 | Comment out unused code | Johannes Ranke | 2 | -27/+27 | |
2019-02-27 | Increase test coverage | Johannes Ranke | 1 | -8/+17 | |
2019-02-26 | NAFTA SOP Attachment vignette | Johannes Ranke | 2 | -3/+7 | |
Also: - Change rounding in print.nafta - Add dots argument to nafta() - Use cores=1 in examples - Restrict N in IORE model to values > 0 | |||||
2019-02-26 | Add another NAFTA dataset + test | Johannes Ranke | 1 | -1/+2 | |
2019-02-25 | Static documentation rebuilt by current pkgdown | Johannes Ranke | 1 | -3/+7 | |
2019-02-22 | Implement the NAFTA evaluation scheme | Johannes Ranke | 1 | -0/+105 | |
2019-02-22 | Further adaptations for the logistic model | Johannes Ranke | 3 | -4/+9 | |
- Make mmkin compatible - Return DT50 values corresponding to k0 and kmax - Turn incompatible parameter names in parms.ini from an error to a warning, in order to make it possible to use this argument in calls to mmkin | |||||
2019-02-21 | Add the logistic model | Johannes Ranke | 6 | -17/+43 | |
2019-01-31 | Restore NAMESPACE | Johannes Ranke | 1 | -16/+1 | |
which was accidentally overwritten by pkgdown -> roxygen | |||||
2019-01-31 | Export to CAKE study file | Johannes Ranke | 1 | -0/+80 | |
2018-11-26 | Add AIC method for mmkin column objects | Johannes Ranke | 1 | -0/+30 | |
Static documentation rebuilt by pkgdown | |||||
2018-11-26 | Support logLik and AIC for manual weights | Johannes Ranke | 1 | -2/+11 | |
Static documentation rebuilt by pkgdown | |||||
2018-11-23 | Add logLik method to enable AIC() on mkinfit models | Johannes Ranke | 2 | -1/+36 | |
Further relax two tests to pass build on Travis | |||||
2018-11-23 | Improve initialisation of mkinds | Johannes Ranke | 1 | -10/+13 | |
Check in work from the beginning of the week | |||||
2018-09-21 | Improve fitting the two-component error model | Johannes Ranke | 1 | -47/+83 | |
with respect to accuracy and robustness. | |||||
2018-09-20 | Use the median absolute deviance for fitting | Johannes Ranke | 1 | -5/+16 | |
as the absolute value is a biased estimator for the standard deviation | |||||
2018-09-18 | Make mkinpredict generic and write a method for fits | Johannes Ranke | 1 | -25/+52 | |
2018-09-18 | Support maximum twa for HS, test, docs | Johannes Ranke | 1 | -2/+16 | |
2018-09-15 | Remove correlation test as gmkin on windows still crashes | Johannes Ranke | 1 | -18/+10 | |
Clean up the code a bit | |||||
2018-09-15 | Respect reps argument in add_err | Johannes Ranke | 1 | -2/+2 | |
2018-09-14 | Update NEWS | Johannes Ranke | 1 | -2/+4 | |
Static documentation rebuilt by pkgdown | |||||
2018-09-14 | Avoid nls crashes by checking for a correlation first | Johannes Ranke | 1 | -0/+7 | |
2018-09-13 | Enable new error model in gmkin | Johannes Ranke | 1 | -7/+20 | |
Also improve the irls fitting of the error model and add a test for FOCUS_2006_C where the second component of the error model is zero | |||||
2018-07-18 | Correct references to the Rocke and Lorenzato model | Johannes Ranke | 3 | -6/+6 | |
Rename 'sigma_rl' to 'sigma_twocomp' as the Rocke and Lorenzato model assumes lognormal distribution for large y. Rebuild static documentation. |