STAGGER

STAGGER concepts stream generator.

This generator is an implementation of the dara stream with abrupt concept drift, as described in ¹.

The STAGGER concepts are boolean functions f with three features describing objects: size (small, medium and large), shape (circle, square and triangle) and colour (red, blue and green).

f options:

1. True if the size is small and the color is red.

2. True if the color is green or the shape is a circle.

3. True if the size is medium or large

Concept drift can be introduced by changing the classification function. This can be done manually or using datasets.synth.ConceptDriftStream.

One important feature is the possibility to balance classes, which means the class distribution will tend to a uniform one.

Parameters

• classification_function

  Type → int

  Default → 0

  Classification functions to use. From 0 to 2.

• seed

  Type → int | None

  Default → None

  Random seed for reproducibility.

• balance_classes

  Type → bool

  Default → False

  Whether to balance classes or not. If balanced, the class distribution will converge to an uniform distribution.

Attributes

• desc

  Return the description from the docstring.

Examples

from river.datasets import synth

dataset = synth.STAGGER(classification_function = 2, seed = 112,
                     balance_classes = False)

for x, y in dataset.take(5):
    print(x, y)

{'size': 1, 'color': 2, 'shape': 2} 1
{'size': 2, 'color': 1, 'shape': 2} 1
{'size': 1, 'color': 1, 'shape': 2} 1
{'size': 0, 'color': 1, 'shape': 0} 0
{'size': 2, 'color': 1, 'shape': 0} 1

Methods

generate_drift

Generate drift by switching the classification function at random.

take

Iterate over the k samples.

Parameters

• k — 'int'

Notes

The sample generation works as follows: The 3 attributes are generated with the random number generator. The classification function defines whether to classify the instance as class 0 or class 1. Finally, data is balanced, if this option is set by the user.

---

1. Schlimmer, J. C., & Granger, R. H. (1986). Incremental learning from noisy data. Machine learning, 1(3), 317-354. ↩