mlshell.plot

The mlshell.plot includes utils to visualize results.

Classes

Plotter()

Visualize results.
class mlshell.plot.Plotter

Bases: object

Visualize results.

Methods

plot(pipeline, metrics, datasets, runs, …)

Plot results.
plot(pipeline, metrics, datasets, runs, validator, logger)

Plot results.

Parameters

  • pipeline (mlshell.Pipeline) – Pipeline.

  • metrics (list of class:mlshell.Metric) – Metrics to evaluate.

  • datasets (dict of class:mlshell.Dataset) – Datasets to evaluate on: {‘dataset_id’: dataset}. For classifier dataset.meta should contains pos_labels_ind key.

  • runs (dict) – Resilts for pipeline-dataset pair: {‘dataset_id’: optimizer.update_best output}

  • validator (mlshell.model_selection.Validator) – Object to evaluate vectorized metrics: validator.validate( pipeline, metrics, datasets, logger, vector=True) .

  • logger (logging.Logger) – Logger.