experimentData-class: Experiment data

In familiar: End-to-End Automated Machine Learning and Model Evaluation

Experiment data

Description

An experimentData object contains information concerning the experiment. These objects can be used to instantiate multiple experiments using the same iterations, feature information and variable importance.

Details

experimentData objects are primarily used to improve reproducibility, since these allow for training models on a shared foundation.

Slots

experiment_setup

Contains regarding the experimental setup that is used to generate the iteration list.

iteration_list

List of iteration data that determines which instances are assigned to training, validation and test sets.

feature_info

Feature information objects. Only available if the experimentData object was generated using the precompute_feature_info or precompute_vimp functions.

vimp_table_list

List of variable importance table objects. Only available if the experimentData object was created using the precompute_vimp function.

project_id

Identifier of the project that generated the experimentData object.

familiar_version

Version of the familiar package used to create this experimentData.

See Also

precompute_data_assignment precompute_feature_info, precompute_vimp

familiar documentation built on Sept. 30, 2024, 9:18 a.m.