Changelog
Source:NEWS.md
dynConfiR 0.1.0 (June 2024)
- change in the precision argument for all models. It is now an integer argument representing the approximate precision of the density in digits. Default is 6 for all models.
- renamed the DDMConf model to DDConf to align with published literature
- renamed the functions for dynaViTE model (ddynaViTE and rdynaViTE);
- the functions dWEV and rWEV will be kept for now but maybe removed in future releases of the package, they will produce a deprecation warning now
- Minor changes:
- Bug fix in fitting Race Models with time-dependend confidence variable. Before, the weight wrt was bound from above by the weight wx because of a bug.
- Added a UserInterrupt() call in all longer-running C-functions.
dynConfiR 0.0.4 (January 2024)
CRAN release: 2024-01-29
Fixed a CRAN note and improved robustness in fitting functions, and simulations. Also improved initial parameter sets for model fitting.
dynConfiR 0.0.3 (April 2023)
CRAN release: 2023-06-27
Improvements
- added the dynaViTE model (generalization of dynWEV with time-dependent confidence variable), which includes the additional lambda which can also be fitted
- improved fitting procedure for dynaViTE/dynWEV and 2DSD: confidence thresholds for starting parameters in the initial grid search are now estimated using simulations; this decreases the size of the initial grid drastically, while improving the initial values
- adapted code to changes in the new dplyr release (1.1.1)
dynConfiR 0.0.2 (December 2022)
CRAN release: 2022-12-07
Improvements
- added the DDMConf model (density:
dDDMConf
, RNG:rDDMConf
), also for fitting and prediction. -
d2DSD
andddynaViTE
added argumentlambda
. DynWEV and 2DSD are generalized with the confidence measure explicitly depending on decision time.lambda
controls the power of decision time by which the confidence measure is divided. -
d2DSD
andddynaViTE
added argumentstop_on_zero
: For the calculation of likelihoods, it is useful to stop as soon as one probability is 0, since then the likelihood is 0. - included starting point and drift rate variation in IRMt (experimental!, see
dIRM2
) - fitting 2DSD and dynWEV: improved the finding of stating values for confidence thresholds.
- many bug fixes and increased robustness