seldonian.dataset.SupervisedMetaData

class SupervisedMetaData(sub_regime, all_col_names, feature_col_names, label_col_names, sensitive_col_names=[])

Bases: MetaData

__init__(sub_regime, all_col_names, feature_col_names, label_col_names, sensitive_col_names=[])

Class for holding supervised learning dataset metadata

Parameters:

  • sub_regime (str) – The sub-category of the machine learning algorithm, e.g., “classification” or “regression”.

  • all_col_names (list(str)) – A list of all of the column names in the dataset, including any sensitive attributes and labels.

  • feature_col_names (list(str)) – A list of all of the feature column names in the dataset.

  • label_col_names (list(str)) – A list of all of the label column names in the dataset.

  • sensitive_col_names (list(str), defaults to None) – A list the sensitive column names in the dataset, if any.

__repr__()

Return repr(self).

Methods