LeveragingBaggingClassifier

Leveraging Bagging ensemble classifier.

Leveraging Bagging [^1] is an improvement over the Oza Bagging algorithm. The bagging performance is leveraged by increasing the re-sampling. It uses a poisson distribution to simulate the re-sampling process. To increase re-sampling it uses a higher w value of the Poisson distribution (agerage number of events), 6 by default, increasing the input space diversity, by attributing a different range of weights to the data samples.

To deal with concept drift, Leveraging Bagging uses the ADWIN algorithm to monitor the performance of each member of the enemble If concept drift is detected, the worst member of the ensemble (based on the error estimation by ADWIN) is replaced by a new (empty) classifier.

Parameters

• model (base.Classifier)

  The classifier to bag.

• n_models (int) – defaults to 10

  The number of models in the ensemble.

• w (float) – defaults to 6

  Indicates the average number of events. This is the lambda parameter of the Poisson distribution used to compute the re-sampling weight.

• adwin_delta (float) – defaults to 0.002

  The delta parameter for the ADWIN change detector.

• bagging_method (str) – defaults to bag

  The bagging method to use. Can be one of the following:
  * 'bag' - Leveraging Bagging using ADWIN.
  * 'me' - Assigns \(weight=1\) if sample is misclassified, otherwise \(weight=error/(1-error)\).
  * 'half' - Use resampling without replacement for half of the instances.
  * 'wt' - Resample without taking out all instances.
  * 'subag' - Resampling without replacement.
  

• seed (int) – defaults to None

  Random number generator seed for reproducibility.

Attributes

• bagging_methods

  Valid bagging_method options.

• models

Examples

>>> from river import datasets
>>> from river import ensemble
>>> from river import evaluate
>>> from river import linear_model
>>> from river import metrics
>>> from river import optim
>>> from river import preprocessing

>>> dataset = datasets.Phishing()

>>> model = ensemble.LeveragingBaggingClassifier(
...     model=(
...         preprocessing.StandardScaler() |
...         linear_model.LogisticRegression()
...     ),
...     n_models=3,
...     seed=42
... )

>>> metric = metrics.F1()

>>> evaluate.progressive_val_score(dataset, model, metric)
F1: 88.73%

Methods

append

S.append(value) -- append value to the end of the sequence

Parameters

• item

clear

S.clear() -> None -- remove all items from S

copy

count

S.count(value) -> integer -- return number of occurrences of value

Parameters

• item

extend

S.extend(iterable) -- extend sequence by appending elements from the iterable

Parameters

• other

index

S.index(value, [start, [stop]]) -> integer -- return first index of value. Raises ValueError if the value is not present.

Supporting start and stop arguments is optional, but recommended.

Parameters

• item
• args

insert

S.insert(index, value) -- insert value before index

Parameters

• i
• item

learn_one

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

Parameters

• x
• y

Returns

self

pop

S.pop([index]) -> item -- remove and return item at index (default last). Raise IndexError if list is empty or index is out of range.

Parameters

• i – defaults to -1

predict_one

Predict the label of a set of features x.

Parameters

• x (dict)

Returns

typing.Union[bool, str, int, NoneType]: The predicted label.

predict_proba_one

Averages the predictions of each classifier.

Parameters

• x

remove

S.remove(value) -- remove first occurrence of value. Raise ValueError if the value is not present.

Parameters

• item

reverse

S.reverse() -- reverse IN PLACE

sort