Used in case when folding is needed, but not the evaluation
Utility used for transparent injection of the evaluator into training chain.
Utility used for transparent injection of the evaluator into training chain. Evaluator is applied only while fitting (it adds an extra summary block), but has no other traces in the final model (does not affect predictions).
Utility used to filter out test data before passing to estimator.
This is a simple workaround to add kind of grouping by test/train column for evaluators without embedded support for grouping (eg.
This is a simple workaround to add kind of grouping by test/train column for evaluators without embedded support for grouping (eg. BinaryClassificationEvaluator).
Adds folds (foldNum column) to the dataset before passing it to the nested estimator.
Adds folds (foldNum column) to the dataset before passing it to the nested estimator.
Nested predictor for fitting the model
Transformer adding folds (by default based on row hash)
Estimator returning a model fit by nested predictor on a dataset with extra foldNum column
Performs a cross validation given predictor and evaluator.
Performs a cross validation given predictor and evaluator. Returns a model with summary blocks extended with foldNum column.
Split into folds is done based on the hash of entire row.
Nested predictor for fitting the model.
Evaluator for creating a metric.
Number of folds for validation (defeult 10)
Number of parallel folds training.
Estimator which returns a model fit by the nested predictor on the entire dataset with summary blocks extended with numFolds column.
Fit and then evaluate model.
Fit and then evaluate model. Results of evaluation is stored into a dedicated summary block.
Used to fit the model
Used to evaluate the model.
Estimator which returns a model fit by nested predictor with extra summary block for metrics, produced by evaluator.
Performs a cross validation given predictor and evaluator.
Performs a cross validation given predictor and evaluator. Returns a model with summary blocks extended with foldNum column.
Split into folds is expected to be done externaly.
Nested predictor for fitting the model.
Evaluator for creating a metric.
Number of folds for validation (defeult 10)
Number of threads to run validation.
Estimator which returns a model fit by the nested predictor on the entire dataset with summary blocks extended with numFolds column.