ADWINConfig

class frouros.detectors.concept_drift.streaming.window_based.ADWINConfig(clock: int = 32, delta: float = 0.002, m: int = 5, min_window_size: int = 5, min_num_instances: int = 10)

ADWIN (ADaptive WINdowing) [bifet2007learning] configuration.

Parameters:

• clock (int) – clock value, default to 32

• delta (float) – confidence value, default to 0.002

• m (int) – controls the amount of memory used and the closeness of the cutpoints checked, default to 5

• min_window_size (int) – minimum numbers of instances per window to start looking for changes, default to 5

• min_num_instances (int) – minimum numbers of instances to start looking for changes, default to 10

References:

[bifet2007learning]

Bifet, Albert, and Ricard Gavalda. “Learning from time-changing data with adaptive windowing.” Proceedings of the 2007 SIAM international conference on data mining. Society for Industrial and Applied Mathematics, 2007.

property clock: int

Clock value property.

Returns:

confidence interval to determine if drift is occurring

Return type:

int

property delta: float

Delta value property.

Returns:

confidence interval to determine if drift is occurring

Return type:

float

property m: int

M value property.

Returns:

controls the amount of memory used and the closeness

of the cutpoints checked :rtype: int

property min_window_size: int

Minimum window size value property.

Returns:

minimum window size value per each window

Return type:

int

property min_num_instances: int

Minimum number of instances property.

Returns:

minimum number of instances to start looking for changes

Return type:

int