SAMkNN

class capymoa.classifier.SAMkNN

Bases: MOAClassifier

Self Adjusted Memory k Nearest Neighbor.

Self Adjusted Memory k Nearest Neighbor (SAMkNN) [1] is a lazy classifier.

>>> from capymoa.classifier import SAMkNN
>>> from capymoa.datasets import ElectricityTiny
>>> from capymoa.evaluation import prequential_evaluation
>>>
>>> stream = ElectricityTiny()
>>> classifier = SAMkNN(stream.get_schema())
>>> results = prequential_evaluation(stream, classifier, max_instances=1000)
>>> print(f"{results['cumulative'].accuracy():.1f}")
78.6

[1]

Losing, V., Hammer, B., & Wersing, H. (2016, December). KNN classifier with self adjusting memory for heterogeneous concept drift. In 2016 IEEE 16th international conference on data mining (ICDM) (pp. 291-300). IEEE.

__init__(

schema: Schema,

random_seed: int = 1,

k: int = 5,

limit: int = 5000,

min_stm_size: int = 50,

relative_ltm_size: float = 0.4,

recalculate_stm_error: bool = False,

)

Self Adjusted Memory k Nearest Neighbor (SAMkNN) Classifier

Parameters:

• schema – The schema of the stream.

• random_seed – The random seed passed to the MOA learner.

• k – The number of nearest neighbors.

• limit – The maximum number of instances to store.

• min_stm_size – The minimum number of instances in the STM.

• relative_ltm_size – The allowed LTM size relative to the total limit.

• recalculate_stm_error – Recalculates the error rate of the STM for size adaption (Costly operation). Otherwise, an approximation is used.

cli_help()

predict(instance)

Predict the label of an instance.

The base implementation calls predict_proba() and returns the label with the highest probability.

Parameters:

instance – The instance to predict the label for.

Returns:

The predicted label or None if the classifier is unable to make a prediction.

predict_proba(instance)

Return probability estimates for each label.

Parameters:

instance – The instance to estimate the probabilities for.

Returns:

An array of probabilities for each label or None if the classifier is unable to make a prediction.

train(instance)

Train the classifier with a labeled instance.

Parameters:

instance – The labeled instance to train the classifier with.

random_seed: int

The random seed for reproducibility.

When implementing a classifier ensure random number generators are seeded.

schema: Schema

The schema representing the instances.