Majority Vote Alerting

The majority vote alerting method is a simple yet effective approach to detect drifts. It works by aggregating the results of multiple detectors or the same detector over sevearl instances and making a decision based on the majority of the votes. This method can help to reduce false positives and improve the overall accuracy of change point detection.

Majorty Vote Alerting Classes

class source.alerting.majority_vote.MajorityVote(threshold: float = 0.5)

Bases: object

Majority voting for change point detection (0 = no change point, 1 = change point).

Parameters:

threshold (float) – Minimum proportion of positive votes required when use_threshold=True. Must be in [0.5, 1.0].

__init__(threshold: float = 0.5)
vote(detections: List[int], use_threshold: bool = False, mode: str = 'hard') → int

Perform majority voting.

Parameters:

  • detections (List[int]) – List of detections (0 or 1).

  • use_threshold (bool) – If True, requires fraction of 1s >= self.threshold.

  • mode (str) – “hard” for binary voting, “soft” for probabilistic voting (values in [0, 1]).

Returns:

1 if change point is detected, else 0.

Return type:

int