Skip to content

BOLEClassifier

Boosting Online Learning Ensemble (BOLE).

A modified version of Oza Online Boosting Algorithm 1. For each incoming observation, each model's learn_one method is called k times where k is sampled from a Poisson distribution of parameter lambda. The first model to be trained will be the one with worst correct_weight / (correct_weight + wrong_weight). The worst model not yet trained will receive lambda values for training from the models that incorrectly classified an instance, and the best model's not yet trained will receive lambda values for training from the models that correctly classified an instance. For more details, see 2.

Parameters

  • model

    Type → base.Classifier

    The classifier to boost.

  • n_models

    Default → 10

    The number of models in the ensemble.

  • seed

    Type → int | None

    Default → None

    Random number generator seed for reproducibility.

  • error_bound

    Default → 0.5

    Error bound percentage for allowing models to vote.

Attributes

  • models

Examples

from river import datasets
from river import ensemble
from river import evaluate
from river import drift
from river import metrics
from river import tree

dataset = datasets.Elec2().take(3000)

model = ensemble.BOLEClassifier(
    model=drift.DriftRetrainingClassifier(
        model=tree.HoeffdingTreeClassifier(),
        drift_detector=drift.binary.DDM()
    ),
    n_models=10,
    seed=42
)

metric = metrics.Accuracy()

evaluate.progressive_val_score(dataset, model, metric)

Accuracy: 93.63%

Methods

learn_one

Update the model with a set of features x and a label y.

Parameters

  • x
  • y
  • kwargs

predict_one

Predict the label of a set of features x.

Parameters

  • x — 'dict'
  • kwargs

Returns

base.typing.ClfTarget | None: The predicted label.

predict_proba_one

Predict the probability of each label for a dictionary of features x.

Parameters

  • x
  • kwargs

Returns

A dictionary that associates a probability which each label.

  1. Oza, N.C., 2005, October. Online bagging and boosting. In 2005 IEEE international conference on systems, man and cybernetics (Vol. 3, pp. 2340-2345). Ieee. ↩

  2. R. S. M. d. Barros, S. Garrido T. de Carvalho Santos and P. M. Gonçalves Júnior, "A Boosting-like Online Learning Ensemble," 2016 International Joint Conference on Neural Networks (IJCNN), 2016, pp. 1871-1878, doi: 10.1109/IJCNN.2016.7727427. ↩