- Basic parallel estimation of all subsets using
`parallel::parSapply`

- Progress bar added for non-parallel sub-model estimation
- Complete dominance reports fraction of total models higher/lower
- Update to
`summary`

method

- Update to
`offset()`

s now allowed with`formula`

method`formula`

input with`.adj`

defunct- Better handing of errors returned by ‘.fct’. Reports the specific error that resulted.
- New
`pkgdown`

-based GitHub page with extra visual design features.

`formula`

method sync with`formula_list`

- Intercept removal and offsets not allowed in
`.all`

and`.set`

`offsets()`

omitted from models with no predictors. Was an issue with`.adj`

’s value

`domir()`

produced error when`.cdl = FALSE`

was combined with`.all`

*lme4*added to suggested list due to reference in documentation and CRAN warning

`formula_list`

method implemented- merger of planned
`list`

and`Formula::Formula`

methods `fmllst2Fml`

can be used to translate`formula_list`

s to`Formula::Formula`

s`offset()`

s allowed in`formula_list`

; to update in`formula`

method

- merger of planned
`formula`

method`.adj`

argument is logical`formula`

s of the form`~ 1`

still accepted but depreciated- other
`formula`

inputs defunct and user pushed to add to`.fct`

- Began linting process for code readability

`formula`

method in`domir()`

and`domin()`

now disallow`offset()`

terms- Were silently removed and not replaced during parsing.

- Argument checks for
`formula`

method produced inconsistent results

- New
`domir`

function- Generic function with three S3 methods:
`formula`

,`Formula`

, and`list`

`Formula`

and`list`

methods under development

`.obj`

submits object on which S3 method dispatch is based`.fct`

submits function that accepts result of same type as`.obj`

and returns a scalar`.set`

,`.all`

and`.adj`

affect`.obj`

entries`.wst`

argument under development (‘within-set’ or Owen decomposition)

- Generic function with three S3 methods:
`domin`

superseded- No longer under active development; bug fixes only

`dominance_scalar`

internal function responsible for computation.`domir`

is front-end to computational engine.- Other computational engines planned. Internal function may be renamed.

`summary.domin`

not recognizing`reverse = TRUE`

.

- removing reference to
`dominanceanalysis`

package not on CRAN currently given request from CRAN maintainers and CRAN note.

- fixing typos and other editorial errors in
*domir_basics*vignette. - Extending conditional dominance discussion with graphic.

- fixing error in email address in
*domir-package*leading to CRAN check warnings.

`conditional`

logical argument to make computing conditional dominance statistics optional- initial working version of
`summary.domin`

- method for computing the strongest dominance designations between independent variables
- update to
`print.domin`

to display strongest dominance designations

- reorganization and update to
*Conceptual Introduction to Dominance Analysis*vignette

- looping used to construct many dominance statistics, replaced with matrix-based methods; intended to simplify and expedite computations

- update to email in
`domin-package`

as requested by CRAN maintainers.

`consmodel`

argument to adjust for baseline fit statistic values`reverse`

argument to change rank and complete dominance interpretation for fit metrics that decrease with better fit*Conceptual Introduction to Dominance Analysis*vignette added

- implemented unit testing framework - covers 82% of
`domin`

- internal use of
`reformulate`

to construct formulas as opposed to`paste`

-ing`+`

-es

`Complete_Dominance`

computation producing inconsistent complete dominance designations - within-order complete dominance checks not capturing all possible model comparisons

- update to format of
`Complete_Dominance`

from integers to logicals - output and warnings no longer suppressed by default - user must silence noisy functions
- minimum of two terms/sets to run
`domin`

(replicates behavior of Stata version) - additional checks (response in formula/orders in formula)
- names of entries in
`domin`

object changed to syntactic R names when not. Affects entries in:`Complete_Dominance`

matrix`Conditional_Dominance`

matrix ## internal

- extensive re-write of computation methods
- pre-allocates container objects to improve performance
- increased functional-ization of subroutines and overall code readability
- many internal functions re-named

- returned value
`.$Subset_Details$Full_Model`

now includes variables in`all`

`Complete_Dominance`

matrix computation- Too many models considered for complete dominance - inconsistent with standard dominance analysis methodology
- Fixed error in suppressing complete dominance (i.e.,
`complete = FALSE`

) resulted in error

`Ensemble_Coordinator`

that submitted factor numbers instead of text in R versions < 4.0

- initial working version of
`domin`

- initial working version of
`print.domin`