R port of the Python module morie.validation: schema validation,
data quality scoring, cross-validation, calibration / discrimination /
decision-curve analysis, overfitting detection, temporal / external
validation, and reproducibility manifests.
Details
Most callables return a named list (class
"morie_validation_result") so the R caller does not need
S4 or R6. Model-fitting routines accept a user-supplied
fit_fn of signature function(X, y) -> model and a
predict_fn of signature function(model, X) -> prob; this
keeps the port framework-agnostic (works with glm, glmnet,
randomForest, xgboost, etc.).