lib.sedna.algorithms.seen_task_learning.task_definition.task_definition

Divide multiple tasks based on data

param samples： Train data:

param see sedna.datasources.BaseDataSource for more detail.:

returns:

  • tasks (All tasks based on training data.)

  • task_extractor (Model with a method to predicting target tasks)

Module Contents

Classes

TaskDefinitionBySVC

Dividing datasets with AgglomerativeClustering based on kernel distance,

TaskDefinitionByDataAttr

Dividing datasets based on the common attributes,
class lib.sedna.algorithms.seen_task_learning.task_definition.task_definition.TaskDefinitionBySVC(**kwargs)[source]

Dividing datasets with AgglomerativeClustering based on kernel distance, Using SVC to fit the clustering result.

Parameters:

None (n_class： int or) – The number of clusters to find, default=2.

__call__(samples: sedna.datasources.BaseDataSource) → Tuple[List[lib.sedna.algorithms.seen_task_learning.artifact.Task], Any, sedna.datasources.BaseDataSource][source]
class lib.sedna.algorithms.seen_task_learning.task_definition.task_definition.TaskDefinitionByDataAttr(**kwargs)[source]

Dividing datasets based on the common attributes, generally used for structured data.

Parameters:

List[Metadata] (attribute：) – metadata is usually a class feature label with a finite values.

__call__(samples: sedna.datasources.BaseDataSource, **kwargs) → Tuple[List[lib.sedna.algorithms.seen_task_learning.artifact.Task], Any, sedna.datasources.BaseDataSource][source]