gaussian_arm

Autoregressive Gaussian monitoring schema for generalized CUSUM.

class pysatl_cpd.algorithms.online.cusum.component.monitoring.gaussian_arm.GaussianARMonitoringSchema[source]

Standardised one-step autoregressive forecast residual.

Computes (predicted - observed) / sqrt(noise_variance) using AR model coefficients and history from the estimating schema.

evaluate(observation, estimates)[source]

Compute standardised AR forecast residual.

Parameters:

observation (ndarray[tuple[int], dtype[double]]) – New observation (must be dim=1).
estimates (EstimatesGaussianAR) – AR estimates containing coefficients, history, intercept, and noise_variance.

Returns:

Standardised residual array of length 1.

Return type:

ndarray[tuple[int], dtype[double]]

Raises:

ValueError – If observation is not dim=1, or if coefficients are not yet fitted.

Reset monitoring schema state.