API Reference / datasets / generators / clustering /

circles.py

Circles generator: two concentric rings.

Produces a small ring nested inside a large one, with Gaussian jitter. The two classes cannot be separated by any straight line, so it is a quick way to show where linear models fail and kernel or density-based methods succeed.

Classes

Circles

class datasets.generators.clustering.circles.Circles(ClusteringGenerator)

Concentric Circles data generator.

Generates two concentric circles (inner and outer). This is a classic non-linearly separable dataset.
Constructor 
__init__(
    self,
    n_samples: int = 100,
    noise: float = 0.05,
    factor: float = 0.5,
    shuffle: bool = True,
    random_state: Optional[int] = None,
)

Parameters

n_samples
int = 100
Number of samples to generate.
noise
float = 0.05
Standard deviation of Gaussian noise.
factor
float = 0.5
Scale factor between inner and outer circle (0 < factor < 1).
shuffle
bool = True
Whether to shuffle the samples.
random_state
int or None = None
Random seed for reproducibility.
python 
>>> from tuiml.datasets.generators.clustering import Circles
>>> gen = Circles(n_samples=1000, noise=0.05, factor=0.5, random_state=0)
>>> data = gen.generate()
>>> data.X.shape
(1000, 2)
>>> sorted(set(data.y.tolist()))          # 0 = outer ring, 1 = inner ring
[0, 1]

Methods

generate (self) -> GeneratedData

Generate concentric circles data.

Returns
GeneratedData
Generated dataset with feature array X of shape (n_samples, 2) and circle labels y (0=outer, 1=inner).
get_parameter_schema (cls) -> Dict[str, Any]

Return JSON Schema for constructor parameters.