Changelog • cognitivemodels

cognitivemodels 0.0.12

Bugfixes

New features

Added an optional argument prior_sum to Bayesian learning models to set and control the sum constraints of the prior (hyper-)parameter, see ?bayes
Added a new model shift_c() that models a change point in time for shifting between two values or two predictions. See ?shift
cpt() now allows weighting = NA and value = NA as arguments, which leads to no probability weighting (treating probabilities as is) or no subjective value transformation (treating outcomes as is) in the model, respectively, see ?cpt

Potentially breaking

In predict.cm changed the order or arguments such that the argument newdata is at the second position to comply with predict.glm methods

Bugfixes

Fixed bugs in ROI optimization
Fixed bug in add_constraints() in cognitivemodel lego variant
Made dropping of choicerule error and warning easier to understand.
Declared solvers() as depreciated with a warning, from now on use cm_solvers() to show the optimization solvers.