BuiltInOp

class hana_ml.algorithms.pal.pipeline.BuiltInOp(op_name, **kwargs)

Pipeline built-in operators.

Parameters
op_name{'OneHotEncoder', 'LabelEncoder', 'CBEncoder', 'PolynomialFeatures', 'TargetEncoder'}
**kwargskeywords arguments

Arbitrary keywords arguments passed to the specified built-in operator specified by op_name.

Valid operator-parameter setting is displayed as follows:

  • OneHotEncoder

    • ignore_unknown : type int, with valid values {0, 1} and default value 1, where 0 means does not ignore the any unknown categorical value in predict data(i.e. throw-error if encountered), and 1 means encoding the unknown categories from -n to -1(n is the number of unknown categories in predict data).

    • minimum_fraction : type float with default value 0.0, which specifies the minimum fraction of uniques values in a feature for it to be considered categorical.

  • LabelEncoder

    • ignore_unknown : type int with valid values {0, 1} and default value 1, where 0 means does not ignore the any unknown categorical value in predict data(i.e. throw-error if encountered), and 1 means encoding the unknown categories from -n to -1(n is the number of unknown categories in predict data).

  • PolynomialFeatures

    • min_degree : type int with default value 1. It specifies the minimum polynomial degree of generated features.

    • max_degree : type int with default value 2. It specifies the maximum polynomial degree of generated features.

    • interaction_only : type bool with default value False. If set as true, only interaction features are produced.

    • include_bias : type bool with default value True. It specifies whether or not to include a column of constant 1 in the generated polynomial features.

  • CBEncoder

    None

Methods

fit(data)

Dummy function.

fit_transform(data)

Dummy function.

fit(data)

Dummy function.

fit_transform(data)

Dummy function.

Inherited Methods from PALBase

Besides those methods mentioned above, the BuiltInOp class also inherits methods from PALBase class, please refer to PAL Base for more details.