Contents
ComparisonFamily
Internal base class for comparison families
ComparisonFamily dispatches the incremental, heterogeneity and contrast
comparisons for associate$compare(). Each subclass fits one outcome family.
Subclasses override validate.context() and prepare.data(), and implement
the three private run.*() methods.
Public fields
family.name — Canonical short name for this comparison family
(e.g. "lm", "glm", "cox").
Methods
Public methods
ComparisonFamily$new()ComparisonFamily$run()ComparisonFamily$validate.context()ComparisonFamily$prepare.data()ComparisonFamily$clone()
Method new()
Store the canonical family key.
Usage
ComparisonFamily$new(family.name)
Arguments
family.name — Canonical short name for this family.
Method run()
Dispatch to the comparison type (incremental,
heterogeneity or contrast). The type comes from fit.specs$type, then
cell$type, and defaults to "incremental". Calls validate.context()
then prepare.data() before routing.
Usage
ComparisonFamily$run(context)
Arguments
context — List with fields frame, cell, fit.specs and
conf.level.
Returns
A PolyGeniusAssociation.
Method validate.context()
Validate the frame before dispatching. Subclasses override this; the base version accepts everything.
Usage
ComparisonFamily$validate.context(frame, cell, fit.specs, type)
Arguments
frame — data.table resolved from the study.
cell — Fit-cell list (outcome, predictor, stratum, etc.).
fit.specs — Extra fitting options from the user call.
type — Comparison type: "incremental", "heterogeneity" or "contrast".
Returns
TRUE, invisibly, or an abort.
Method prepare.data()
Extract family-specific clean data from the resolved frame. Subclasses override this; the base version aborts.
Usage
ComparisonFamily$prepare.data(frame, cell, fit.specs)
Arguments
frame — data.table resolved from the study.
cell — Fit-cell list.
fit.specs — Extra fitting options.
Returns
A list of ready-to-use fit inputs.
Method clone()
The objects of this class are cloneable with this method.
Usage
ComparisonFamily$clone(deep = FALSE)
Arguments
deep — Whether to make a deep clone.
See Also
associate$compare, RegressionFamily
Other comparison-families:
CompetingRiskComparison,
CoxComparison,
LinearComparison,
LogisticComparison