pdmlabs.evaluation.default_evaluators

pdmlabs.evaluation.default_evaluators#

Classes

BaseADEvaluator([debug])

Base class for AD-style metrics.

DefaultADEvaluator([debug])

DefaultClassificationEvaluator([debug])

DefaultRULEvaluator([debug])

DefaultSurvEvaluator([debug])

class pdmlabs.evaluation.default_evaluators.BaseADEvaluator(debug=False)#

Bases: EvaluatorInterface

Base class for AD-style metrics. Computes standard PdM metrics like AD1_AUC, VUS, etc.

class pdmlabs.evaluation.default_evaluators.DefaultADEvaluator(debug=False)#

Bases: BaseADEvaluator

evaluate(experiment, **kwargs) → dict#

Compute evaluation metrics and return them.

Parameters:

  • experiment – The PdMExperiment instance running the evaluation.

  • **kwargs – Additional data required for evaluation depending on the experiment flavor. Common kwargs include: - result_scores: The predicted scores/values. - result_dates: The timestamps corresponding to the predictions. - results_isfailure / result_labels: Ground-truth labels or failure indicators. - plot_dictionary: A dictionary to store plotting data if debug=True. - rtfs: Run-to-failure indicators (for RUL/SA).

Returns:

A dictionary mapping metric names (str) to their computed values.

Example: {‘my_custom_f1’: 0.85, ‘my_custom_auc’: 0.92}

Return type:

dict

class pdmlabs.evaluation.default_evaluators.DefaultClassificationEvaluator(debug=False)#

Bases: BaseADEvaluator

evaluate(experiment, **kwargs) → dict#

Compute evaluation metrics and return them.

Parameters:

  • experiment – The PdMExperiment instance running the evaluation.

  • **kwargs – Additional data required for evaluation depending on the experiment flavor. Common kwargs include: - result_scores: The predicted scores/values. - result_dates: The timestamps corresponding to the predictions. - results_isfailure / result_labels: Ground-truth labels or failure indicators. - plot_dictionary: A dictionary to store plotting data if debug=True. - rtfs: Run-to-failure indicators (for RUL/SA).

Returns:

A dictionary mapping metric names (str) to their computed values.

Example: {‘my_custom_f1’: 0.85, ‘my_custom_auc’: 0.92}

Return type:

dict

class pdmlabs.evaluation.default_evaluators.DefaultRULEvaluator(debug=False)#

Bases: EvaluatorInterface

evaluate(experiment, **kwargs) → dict#

Compute evaluation metrics and return them.

Parameters:

  • experiment – The PdMExperiment instance running the evaluation.

  • **kwargs – Additional data required for evaluation depending on the experiment flavor. Common kwargs include: - result_scores: The predicted scores/values. - result_dates: The timestamps corresponding to the predictions. - results_isfailure / result_labels: Ground-truth labels or failure indicators. - plot_dictionary: A dictionary to store plotting data if debug=True. - rtfs: Run-to-failure indicators (for RUL/SA).

Returns:

A dictionary mapping metric names (str) to their computed values.

Example: {‘my_custom_f1’: 0.85, ‘my_custom_auc’: 0.92}

Return type:

dict

class pdmlabs.evaluation.default_evaluators.DefaultSurvEvaluator(debug=False)#

Bases: EvaluatorInterface

evaluate(experiment, **kwargs) → dict#

Compute evaluation metrics and return them.

Parameters:

  • experiment – The PdMExperiment instance running the evaluation.

  • **kwargs – Additional data required for evaluation depending on the experiment flavor. Common kwargs include: - result_scores: The predicted scores/values. - result_dates: The timestamps corresponding to the predictions. - results_isfailure / result_labels: Ground-truth labels or failure indicators. - plot_dictionary: A dictionary to store plotting data if debug=True. - rtfs: Run-to-failure indicators (for RUL/SA).

Returns:

A dictionary mapping metric names (str) to their computed values.

Example: {‘my_custom_f1’: 0.85, ‘my_custom_auc’: 0.92}

Return type:

dict

Contents