ZERO

ZERO returns an all-zero score and serves as a negative-control baseline. A metric or benchmark on which ZERO performs well deserves careful inspection.

Parameters

This detector has no parameters.

Methods

fit(time_series)

No-op included for interface compatibility.

transform(time_series)

time_series : numpy.ndarray

Input whose shape and dtype are copied.

Returns

score : numpy.ndarray

Zeros with the same shape and dtype as time_series.

Minimal Example

import numpy as np
from changepoynt.algorithms.baseline import ZERO

signal = np.array([0.2, 0.4, 1.8, 2.0])
score = ZERO().transform(signal)

Background: van den Burg and Williams, An Evaluation of Change Point Detection Algorithms.