econml.validate.UpliftEvaluationResults

class econml.validate.UpliftEvaluationResults(params: List[float], errs: List[float], pvals: List[float], treatments: numpy.array, curve_data_dict: Dict[Any, pandas.core.frame.DataFrame])[source]

Bases: object

Results class for uplift curve-based tests.

Parameters
  • params (list or numpy array of floats) – Sequence of estimated QINI coefficient values

  • errs (list or numpy array of floats) – Sequence of estimated QINI coefficient standard errors

  • pvals (list or numpy array of floats) – Sequence of estimated QINI coefficient p-values

  • treatments (list or numpy array of floats) – Sequence of treatment labels

  • curve_data_dict (dict) – Dictionary mapping treatment levels to dataframes containing necessary data for plotting uplift curves

__init__(params: List[float], errs: List[float], pvals: List[float], treatments: numpy.array, curve_data_dict: Dict[Any, pandas.core.frame.DataFrame])[source]

Methods

__init__(params, errs, pvals, treatments, ...)

plot_uplift(tmt[, err_type])

Plots uplift curves.

summary()

Constructs dataframe summarizing the results of the QINI test.

plot_uplift(tmt: Any, err_type: Optional[str] = None)[source]

Plots uplift curves.

Parameters
  • tmt (any (sortable)) – Name of treatment to plot.

  • err_type (str) – Type of error to plot. Accepted values are normal (None), two-sided uniform confidence band (‘ucb2’), or 1-sided uniform confidence band (‘ucb1’).

Return type

matplotlib plot with percentage treated on x-axis and uplift metric (and 95% CI) on y-axis

summary()[source]

Constructs dataframe summarizing the results of the QINI test.

Parameters

None

Return type

pandas dataframe containing summary of QINI test results