nannyml.drift.model_outputs.univariate.statistical.calculator module

Calculates drift for model predictions and model outputs using statistical tests.

class nannyml.drift.model_outputs.univariate.statistical.calculator.StatisticalOutputDriftCalculator(y_pred: str, timestamp_column_name: str, problem_type: Union[str, ProblemType], y_pred_proba: Optional[Union[str, Dict[str, str]]] = None, chunk_size: Optional[int] = None, chunk_number: Optional[int] = None, chunk_period: Optional[str] = None, chunker: Optional[Chunker] = None)[source]

Bases: AbstractCalculator

Calculates drift for model predictions and model outputs using statistical tests.

Creates a new StatisticalOutputDriftCalculator.

Parameters:

  • y_pred_proba (ModelOutputsType) – Name(s) of the column(s) containing your model output. Pass a single string when there is only a single model output column, e.g. in binary classification cases. Pass a dictionary when working with multiple output columns, e.g. in multiclass classification cases. The dictionary maps a class/label string to the column name containing model outputs for that class/label.

  • y_pred (str) – The name of the column containing your model predictions.

  • timestamp_column_name (str) – The name of the column containing the timestamp of the model prediction.

  • chunk_size (int, default=None) – Splits the data into chunks containing chunks_size observations. Only one of chunk_size, chunk_number or chunk_period should be given.

  • chunk_number (int, default=None) – Splits the data into chunk_number pieces. Only one of chunk_size, chunk_number or chunk_period should be given.

  • chunk_period (str, default=None) – Splits the data according to the given period. Only one of chunk_size, chunk_number or chunk_period should be given.

  • chunker (Chunker, default=None) – The Chunker used to split the data sets into a lists of chunks.

Examples

>>> import nannyml as nml
>>>
>>> reference_df, analysis_df, _ = nml.load_synthetic_binary_classification_dataset()
>>>
>>> calc = nml.StatisticalOutputDriftCalculator(
>>>     y_pred_proba='y_pred_proba',
>>>     y_pred='y_pred',
>>>     timestamp_column_name='timestamp'
>>> )
>>> calc.fit(reference_df)
>>> results = calc.calculate(analysis_df)
>>>
>>> print(results.data)  # check the numbers
             key  start_index  ...  y_pred_proba_alert y_pred_proba_threshold
0       [0:4999]            0  ...                True                   0.05
1    [5000:9999]         5000  ...               False                   0.05
2  [10000:14999]        10000  ...               False                   0.05
3  [15000:19999]        15000  ...               False                   0.05
4  [20000:24999]        20000  ...               False                   0.05
5  [25000:29999]        25000  ...                True                   0.05
6  [30000:34999]        30000  ...                True                   0.05
7  [35000:39999]        35000  ...                True                   0.05
8  [40000:44999]        40000  ...                True                   0.05
9  [45000:49999]        45000  ...                True                   0.05
>>>
>>> results.plot(kind='score_drift', metric='p_value', plot_reference=True).show()
>>> results.plot(kind='score_distribution', plot_reference=True).show()
>>> results.plot(kind='prediction_drift', plot_reference=True).show()
>>> results.plot(kind='prediction_distribution', plot_reference=True).show()