Assembled-OpenML Classes

class assembledopenml.openml_assembler.OpenMLAssembler(openml_metric_name='area_under_roc_curve', maximize_metric=True, nr_base_models=50)

rebuild(openml_task_id, valid_predictors)

Rebuild a metatask from a list of predictors

Parameters:

• openml_task_id (int) – An Task ID from OpenML for which to build a metatask.

• valid_predictors (List[str]) – A list of predictor names (in the metatask format) which shall be part of the metatask

Returns:

meta_task – A Metatask created based on the collector’s settings for the OpenML Task provided as input

Return type:

MetaTask

run(openml_task_id)

Search through OpenML for valid runs and their predictions.

Parameters:

openml_task_id (int) – An Task ID from OpenML for which to build a metatask.

Returns:

meta_task – A Metatask created based on the collector’s settings for the OpenML Task provided as input

Return type:

MetaTask

class assembledopenml.metaflow.MetaFlow(flow_id, description, run_performance, run_id, conf_tol=0.01)

get_predictions_data(class_labels)

Fill the metaflow object with predictions data (and other relevant variables)

Parameters:

class_labels (List[str]) – The names of each class label