PolyGenius
Contents

LinearRegression

Gaussian linear-regression family for associate$regression()

Fits Gaussian linear models with stats::lm() for continuous outcomes. associate$regression instantiates this family for model = "lm", which is also what model = "auto" falls back to when the outcome is neither a survival descriptor nor binary.

Details

The fitted model is outcome ~ predictor [* interaction] [+ covariates] on complete cases, with case weights when the resolved frame carries a weights column. Covariate adjustment, split.by strata and one predictor-by-modifier interaction are all handled through the engine fit grid rather than inside this class.

Result rows follow the "lm" association schema (schemaLm() in associate-schemas.R): one row per predictor main and interaction coefficient, on the identity scale (effect.scale = "identity", test = "t"), with Wald limits from the residual-df t distribution and n from stats::nobs().

The single artifact is prediction.grid: fitted values with standard errors across the observed predictor range -- grid.resolution equally spaced points for a numeric predictor, the observed levels otherwise -- holding covariates at their reference values.

Super class

PolyGenius::RegressionFamily -> LinearRegression

Methods

Public methods

  • LinearRegression$new()
  • LinearRegression$prepare.frame()
  • LinearRegression$describe.model.terms()
  • LinearRegression$build.formula()
  • LinearRegression$fit.model()
  • LinearRegression$build.summary.rows()
  • LinearRegression$build.artifacts()
  • LinearRegression$clone()

Method new()

Register this family under the key "lm".

Usage

LinearRegression$new()

Returns

A new LinearRegression.

Method prepare.frame()

Reduce the resolved frame to the columns this fit uses and drop rows missing any of them.

Usage

LinearRegression$prepare.frame(frame, cell, fit.specs, conf.level)

Arguments

frame — A data.table resolved from a PolyGeniusStudy, whose roles() tags identify the outcome, predictor, interaction, covariate and weight columns.

cell — Named list describing one fit cell. Unused; columns come from the frame's roles.

fit.specs — Named list of extra fitting options. Unused.

conf.level — Numeric scalar in (0, 1). Unused.

Returns

Named list: data (complete-case data.frame), outcome.name, predictor.name, predictor.label, interaction.name, covariate.names, weights.name (NA_character_ when none) and row.id, the retained row indices into frame.

Method describe.model.terms()

Extend the shared right-hand-side term descriptors with the continuous outcome, whose estimability fails when it has no variance.

Usage

LinearRegression$describe.model.terms(fit.data)

Arguments

fit.data — Named list returned by prepare.frame().

Returns

List of list(column, role, label) descriptors: the base class's terms plus one with role = "outcome.continuous".

Method build.formula()

Build outcome ~ predictor [* interaction] [+ covariates] from the original column names.

Usage

LinearRegression$build.formula(frame)

Arguments

frame — Named list returned by prepare.frame().

Returns

A formula, carrying its printable form as the "display.formula" attribute.

Method fit.model()

Fit the model with stats::lm(), adding case weights when prepare.frame() resolved a weights column.

Usage

LinearRegression$fit.model(formula, frame, ...)

Arguments

formula — A formula from build.formula().

frame — Named list returned by prepare.frame().

... — Further arguments passed to stats::lm().

Returns

A fitted lm object.

Method build.summary.rows()

Turn the fitted model's predictor coefficients into standard result rows on the identity scale.

Usage

LinearRegression$build.summary.rows(
  fit,
  fit.data,
  cell,
  fit.specs,
  conf.level,
  formula
)

Arguments

fit — A fitted lm object.

fit.data — Named list returned by prepare.frame().

cell — Named list describing one fit cell: fit.id, analysis.id, family, outcome, predictor, interaction, stratum.

fit.specs — Named list of extra fitting options. Unused.

conf.level — Numeric scalar in (0, 1). Confidence level for lower/upper, taken from the residual-df t distribution.

formula — Character scalar. Display formula recorded on each row.

Returns

A data.table of "lm"-schema rows, one per predictor main and interaction coefficient. Aborts when no coefficient name matches the predictor.

Method build.artifacts()

Build the fitted-value grid that effect plots read.

Usage

LinearRegression$build.artifacts(
  fit,
  fit.data,
  cell,
  fit.specs,
  conf.level,
  formula
)

Arguments

fit — A fitted lm object.

fit.data — Named list returned by prepare.frame().

cell — Named list describing one fit cell.

fit.specs — Named list of extra fitting options. Unused.

conf.level — Numeric scalar in (0, 1). Unused.

formula — Character scalar. Unused; present for interface parity.

Returns

Named list with one element, prediction.grid: a data.frame of predictor.value, fitted.value and se alongside the fit identity columns (analysis.id, fit.id, schema, family, outcome, predictor, stratum).

Method clone()

The objects of this class are cloneable with this method.

Usage

LinearRegression$clone(deep = FALSE)

Arguments

deep — Whether to make a deep clone.

Examples

# Continuous outcome; this is what model = "auto" resolves to here
associate$regression(data, outcomes = bmi, predictors = PRS_BMI,
                     covariates = c(age, sex))

# Forced, for a 0/1 outcome that would otherwise resolve to "glm"
associate$regression(data, outcomes = affected, predictors = PRS_AD,
                     model = "lm")

See Also

associate$regression, LogisticRegression for binary outcomes, CoxRegression and CompetingRiskRegression for survival outcomes.

Other regression-families: CompetingRiskRegression, CoxRegression, KaplanMeierRegression, LogisticRegression, RegressionFamily