Presentation table for SCM search results

Returns a plain data.frame for an scmSearchResult (or an archived run folder), shaped for a table formatter. Default print() / summary() output stays compact.

Usage

scmSearchTable(x, full = FALSE)

Arguments

x

An scmSearchResult from stepwiseSearch or shotgunSearch, or a path to an archived run folder.

full

Logical. Include the degrees-of-freedom column (df, stepwise only). Default FALSE.

Value

A data.frame.

Details

Stepwise reads StepwiseDetails.csv. Chosen becomes "Y" / "N", and the value columns are named from the Criteria column (Base -2LL, New AIC, ...); mixed or missing criteria keep the engine names BaseValue / NewValue / dValue. Model is the effect tested in that row (see Details).

Shotgun reads the archived scenario index merged with Overall.csv, exposing CovEffects (the enabled fixed effects) next to the fit criteria.

Certara.RsNLME does not import a table formatter; pass the result to one, e.g. flextable::flextable(tab).

For an archived result the table comes from the archive files, so subsetting x first does not reduce the rows.

CovEffects in StepwiseDetails.csv is the cumulative enabled set for a candidate, not the single effect under test, so Model is recovered as a set difference within each Step: the step's base set is the intersection of its candidates when adding and their union when deleting. Steps with a single candidate or an unrecognised direction keep the full CovEffects string.

See also

stepwiseSearch, shotgunSearch, print.scmSearchResult, summary.scmSearchResult

Examples

if (FALSE) { # \dontrun{
tab <- scmSearchTable(stepwiseResult)
flextable::flextable(tab)

scmSearchTable(shotgunResult)
scmSearchTable(attr(stepwiseResult, "searchRunDir"))
} # }