pjml.data.evaluation.split.Split¶
-
class
pjml.data.evaluation.split.Split(split_type: str = 'holdout', partitions: int = 2, partition: int = 0, test_size: float = 0.3, seed: int = 0, fields: str = 'X, Y', **kwargs)[source]¶ Split a given Data field into training/apply set and testing/use set.
Developer: new metrics can be added just following the pattern ‘_fun_xxxxx’ where xxxxx is the name of the new metric.
Parameters: - train_indexes
Indexes of rows to get from data objects during apply().
- test_indexes
Indexes of rows to get from data objects during use().
- fields
Name of the matrices to be modified.
-
__init__(split_type: str = 'holdout', partitions: int = 2, partition: int = 0, test_size: float = 0.3, seed: int = 0, fields: str = 'X, Y', **kwargs)[source]¶ Initialize self. See help(type(self)) for accurate signature.
Methods
__init__(split_type, partitions, partition, …)Initialize self. default_config()Create a copy of the component default configuration. disable_pretty_printing()Disable the pretty-printing. dual_transform(train, …)enable_pretty_printing()Enable the pretty-printing. updated(**kwargs)Clone this component, optionally replacing given params. Attributes
cfserializedcfuuidUUID excluding ‘model’ and ‘enhance’ flags. componentcs1Convert component into a config space with a single component inside it. enhanceridShort uuID First 8 chars of uuid, usually for printing purposes. jsonablelongnamemodelnamepathpretty_printingserializedsidShort uuID First 6 chars of uuid, usually for printing purposes. unwrapSubpipeline inside the first Wrap(). uuidLazily calculated unique identifier for this dataset. wrappedSame as unwrap(), but with the external container Wrap. -
classmethod
default_config() → Dict[str, Any][source]¶ Create a copy of the component default configuration.
Returns: - dict
Copy of the component default configuration.
-
id[source]¶ Short uuID First 8 chars of uuid, usually for printing purposes. First collision expect after 12671943 combinations. :return:
-
unwrap[source]¶ Subpipeline inside the first Wrap().
Hopefully there is only one Wrap in the pipeline. This method performs a depth-first search.
Examples
>>> pipe = Pipeline( >>> File(name='iris.arff'), >>> Wrap(Std(), SVMC()), >>> Metric(function='accuracy') >>> ) >>> pipe.unwrap # -> Chain(Std(), SVMC())
-
updated(**kwargs)[source]¶ Clone this component, optionally replacing given params.
Returns: - A ready to use component.
-
uuid¶ Lazily calculated unique identifier for this dataset.
Should be accessed direct as a class member: ‘uuid’.
Returns: - A unique identifier UUID object.