Fit a scoring pipeline on a training data frame — fit_scoring

Records the feature schema and per-feature training ranges, then calls fit_model(x, y) on the (optionally imputed and scaled) numeric matrix. Imputation and scaling use column means / standard deviations learned on x.

Usage

fit_scoring_pipeline(
  x,
  y,
  fit_model,
  feature_names = names(x),
  impute = FALSE,
  scale = FALSE,
  metadata = list(),
  strict = TRUE
)

Arguments

x: Training data frame (numeric feature columns).
y: Outcome vector.
fit_model: function(x, y) returning a function(x) that gives the positive-class probability.
feature_names: Columns of x to use (default: all, in order).
impute: If TRUE, missing values are replaced by the training column mean.
scale: If TRUE, features are standardised with the training mean and standard deviation.
metadata: Free-form list stored with the bundle (cohort definition, git sha, training dates, ...).
strict: If TRUE (default) columns not in the schema raise an error instead of being dropped.

Value

An omop_scoring_pipeline.