Released on CRAN: September 12, 2024
This is mainly a maintenance release.
Added support: goric now accepts objects of class lm, glm, aov, numeric, lavaan, CTmeta, rma, nlmerMod, glmerMod, and lmerMod
A vignette is added for the benchmark function.
Added support: the benchmark function can be parallelized using the future package.
Added plot() function for benchmark.
Released on CRAN: July 6, 2024
A new function benchmark(): The 'benchmark' function performs benchmarking for models using the Generalized Order-Restricted Information Criterion (Approximation) (GORIC(A)). It can benchmark ratio-of-GORIC(A)-weights, ratio-of-likelihood-weights, GORIC(A) weights, and likihood differences.
Released on CRAN: May 25, 2024
Added support for truncated normal sampling for calculating mixing weights (a.k.a., chi-bar-square weights) to ensure for example that correlation values stay within the allowable range of -1 to 1.
Implemented iterative mixing-weight updating with convergence checking. If mix_weights = "boot", convergence is checked by adding chunks of say 5000 samples per iteration, comparing absolute weight differences to the convergence criterion, and continuing until convergence or the maximum number of iterations is reached.
Fix bug if contraint syntax contained abs(), for example 'x1 < 0 & abs(x2) < 1'. This resulted in a wrong constraint matrix.
Goric() now treats range restrictions as equalities, and Amat is checked for linear dependence, reducing it to row echelon form for penalty calculation. This reduces switching to the bootstrap method for calculating the mixing weights.
Released on CRAN: October 4, 2023
Mainly a maintenance release.
The norm2 package has been replaced by the norm package.
Improved error messages for the goric function.
Released on CRAN: July 7, 2023
A new function evSyn(): GORIC(A) evidence synthesis aggregates the evidence for theory-based hypotheses from multiple studies that may use diverse designs to investigate the same central theory.