FeatureType Class

Defines names of feature types that are recognized for feature engineering in AutoML.

In typical use cases, you use FeatureType attributes for customizing featuration with the AutoMLConfig class and the featurization parameter.

Inheritance

builtins.object

FeatureType

Constructor

FeatureType()

Remarks

FeatureType attributes are used when customizing featurization. For example, to update a column type, use the FeaturizationConfig class as shown in the example.

   featurization_config = FeaturizationConfig()
   featurization_config.add_column_purpose('column1', 'Numeric')
   featurization_config.add_column_purpose('column2', 'CategoricalHash')

For more information, see Configure automated ML experiments.

Attributes

AllNan

AllNan = 'AllNan'

Categorical

Categorical = 'Categorical'

CategoricalHash

CategoricalHash = 'CategoricalHash'

DROP_SET

DROP_SET = {'AllNan', 'Hashes', 'Ignore'}

DateTime

DateTime = 'DateTime'

FULL_SET

FULL_SET = {'AllNan', 'Categorical', 'CategoricalHash', 'DateTime', 'Hashes', 'Ignore', 'Numeric', 'Text'}

Hashes

Hashes = 'Hashes'

Ignore

Ignore = 'Ignore'

Numeric

Numeric = 'Numeric'

Text

Text = 'Text'