base

Base classes for CapyMOA learners.

This module defines the abstract interfaces that all CapyMOA learners implement, including classifiers, regressors, anomaly detectors, clusterers, prediction interval learners, and their semi-supervised and MOA-backed variants.

Classes

AnomalyDetector	Abstract base class for anomaly detector.
Batch	Base class for batch processing in CapyMOA
BatchClassifier	Base class for classifiers that support mini-batches.
BatchRegressor	Base class for regressor that support mini-batches.
Classifier	Base class for classifiers.
ClassifierSSL	Base class for semi-supervised learning classifiers.
Clusterer
ClusteringResult	Abstract clustering result class that has the structure of clusters: centers, weights, radii, and ids.
LearnerParamsMixin	Mixin enabling capturing and reconstructing learner hyper-parameters.
LearnerSpec	Serialized learner specification used for capturing and reconstructing learner hyper-parameters.
MOAAnomalyDetector
MOAClassifier	A wrapper class for using MOA (Massive Online Analysis) classifiers in CapyMOA.
MOAClassifierSSL	Wrapper for using MOA semi-supervised learning classifiers.
MOAClusterer	A wrapper class for using MOA (Massive Online Analysis) clusterers in CapyMOA.
MOAPredictionIntervalLearner
MOARegressor
PredictionIntervalLearner
Regressor
SKClassifier	A wrapper class for using scikit-learn classifiers in CapyMOA.
SKRegressor	A wrapper class for using scikit-learn regressors in CapyMOA.

Functions

learner_from_params

Construct a learner from a serialized fully qualified learner spec.

capymoa.base.learner_from_params(

spec: LearnerSpec,

schema: Any,

random_seed: int = 1,

) → Any

Construct a learner from a serialized fully qualified learner spec.

See LearnerParamsMixin for the parameter capture and serialization support used by learners.