osmose.calibration.uq.emulatorΒΆ

Single-output GP emulator for one OSMOSE output stat.

Builds its OWN sklearn GP β€” it never touches the shared SurrogateCalibrator (the UI and find_optimum depend on that untouched). ARD Matern(2.5) (per-dimension length scales), per-point heteroscedastic noise via alpha, normalize_y=True, fit on natural-log outputs.

Classes

GPEmulator([n_restarts_optimizer, random_state])

One GP emulating a single natural-log output stat over sampling-space X.