mesmer.calibrate_mesmer.train_lv_find_localized_ecov¶
- mesmer.calibrate_mesmer.train_lv_find_localized_ecov(y, wgt_scen_eq, aux, cfg)¶
Find suitable localization radius for empirical covariance matrix and derive localized empirical cov matrix.
- Parameters
y (np.ndarray) – 2d array (sample, gp) of specific target
wgt_scen_eq (np.ndarray) – 1d array (sample) of sample weights
aux (dict) – provides auxiliary variables needed for lv method at hand
[“phi_gc”] (dict with localisation radii as keys and each containing a 2d array (gp, gp) of of Gaspari-Cohn correlation matrix
cfg (module) – config file containing metadata
- Returns
L_sel (numpy.int64) – selected localization radius
ecov (np.ndarray) – 2d empirical covariance matrix array (gp, gp)
loc_ecov (np.ndarray) – 2d localized empirical covariance matrix array (gp, gp)
Notes
Function could also handle determining ecov of several variables but would all have to be passed in same 2d y array (with corresponding wgt_scen_eq, aux[“phi_gc”] shapes)