skactiveml.stream.budgetmanager.BalancedIncrementalQuantileFilter

class skactiveml.stream.budgetmanager.BalancedIncrementalQuantileFilter(w=100, w_tol=50, budget=None)

Bases: BudgetManager

The Balanced Incremental Quantile Filter has been proposed together with Probabilistic Active Learning for Datastreams [1]. It assesses whether a given spatial utility (i.e., obtained via ProbabilisticAL) warrants to query the label in question. The spatial ultilities are compared against a threshold that is derived from a quantile (budget) of the last w observed utilities. To balance the number of queries, w_tol is used to increase or decrease the threshold based on the number of available acquisitions.

Parameters

wint, optional (default=100)

The number of observed utilities that are used to infer the threshold. w should be higher than 0.

w_tolint, optional (default=50)

The window in which the number of acquisitions should stay within the budget. w_tol should be higher than 0.

budgetfloat, optional (default=None)

Specifies the ratio of instances which are allowed to be queried, with 0 <= budget <= 1. See Also BudgetManager.

References

[1] Kottke D., Krempl G., Spiliopoulou M. (2015) Probabilistic Active

Learning in Datastreams. In: Fromont E., De Bie T., van Leeuwen M. (eds) Advances in Intelligent Data Analysis XIV. IDA 2015. Lecture Notes in Computer Science, vol 9385. Springer, Cham.

Methods

get_metadata_routing()	Get metadata routing of this object.
get_params([deep])	Get parameters for this estimator.
query_by_utility(utilities)	Ask the budget manager which utilities are sufficient to query the corresponding instance.
set_params(**params)	Set the parameters of this estimator.
update(candidates, queried_indices, utilities)	Updates the budget manager.

get_metadata_routing()

Get metadata routing of this object.

Please check User Guide on how the routing mechanism works.

Returns

routingMetadataRequest

A MetadataRequest encapsulating routing information.

get_params(deep=True)

Get parameters for this estimator.

Parameters

deepbool, default=True

If True, will return the parameters for this estimator and contained subobjects that are estimators.

Returns

paramsdict

Parameter names mapped to their values.

query_by_utility(utilities)

Ask the budget manager which utilities are sufficient to query the corresponding instance.

Parameters

utilitiesndarray of shape (n_samples,)

The utilities provided by the stream-based active learning strategy, which are used to determine whether sampling an instance is worth it given the budgeting constraint.

Returns

queried_indicesndarray of shape (n_queried_instances,)

The indices of instances represented by utilities which should be queried, with 0 <= n_queried_instances <= n_samples.

set_params(**params)

Set the parameters of this estimator.

The method works on simple estimators as well as on nested objects (such as Pipeline). The latter have parameters of the form <component>__<parameter> so that it’s possible to update each component of a nested object.

Parameters

**paramsdict

Estimator parameters.

Returns

selfestimator instance

Estimator instance.

update(candidates, queried_indices, utilities)

Updates the budget manager.

Parameters

candidates{array-like, sparse matrix} of shape

(n_samples, n_features)

The instances which could be queried. Sparse matrices are accepted only if they are supported by the base query strategy.

queried_indicesarray-like

Indicates which instances from candidates have been queried.

utilitiesndarray of shape (n_samples,)

The utilities based on the query strategy.

Returns

selfEstimatedBudget

The EstimatedBudget returns itself, after it is updated.