pod.compute_pod_curve

compute_pod_curve(
    X_eval,
    mean_model,
    X,
    residuals,
    bandwidth,
    dist_info,
    threshold,
)

Calculates the Probability of Detection (PoD) curve.

Combines the Mean Model, Variance Model, and Error Distribution to compute the probability that the signal exceeds the threshold at every point in X_eval.

Parameters

Name	Type	Description	Default
X_eval	np.ndarray	The grid points to calculate PoD for.	required
mean_model	Any	The fitted sklearn mean response model.	required
X	np.ndarray	Original input data (needed for variance prediction).	required
residuals	np.ndarray	Original residuals (needed for variance prediction).	required
bandwidth	float	Smoothing bandwidth.	required
dist_info	Tuple[str, Tuple]	The (name, params) of the error distribution.	required
threshold	float	The detection threshold value.	required

Returns

Name	Type	Description
	Tuple[np.ndarray, np.ndarray]	Tuple[np.ndarray, np.ndarray]: - pod_curve: Array of probabilities [0, 1] for each point in X_eval. - mean_curve: Array of mean signal response values for X_eval.

Examples

# Assuming we have fitted models (mean_model) and data (X, residuals)
# Calculate the PoD curve for a threshold of 0.5
pod, mean_resp = compute_pod_curve(
    X_eval=np.linspace(0, 10, 100),
    mean_model=mean_model,
    X=X,
    residuals=residuals,
    bandwidth=1.5,
    dist_info=('norm', (0, 1)),
    threshold=0.5
)