CountSampler Class

Default sampler.

Create default sampler.

Inheritance

AbstractSampler

CountSampler

Constructor

CountSampler(seed: int, min_examples_per_class: int = 2000, max_rows: int = 10000, is_constraint_driven: bool = True, task: str = 'classification', train_frac: float | None = None, *args: Any, **kwargs: Any)

Parameters

Name	Description
seed Required	Random seed to use to sample.
min_examples_per_class	Minimum examples per class to sample. default value: 2000
max_rows	Maximum rows to output. default value: 10000
is_constraint_driven	Is constraint driven or not. default value: True
task	The task type corresponding to this sweeping experiment. default value: classification
train_frac	Fraction of data to be considered for constrained training. default value: None

Methods

sample

Sample the give input data.

sample

Sample the give input data.

sample(X: ndarray | DataFrame | spmatrix, y: ndarray | Series | Categorical, cols: str | List[str] | None = None) -> Tuple[ndarray | DataFrame | spmatrix, ndarray | Series | Categorical, SplittingConfig]

Parameters

Name	Description
X Required	Input data.
y Required	Output label.
cols	Optionally, the specific input column(s) over which to sample. default value: None

Returns

Type	Description
	Sampled data.