bundles / scipy 1.18.0 / scipy / interpolate / _bsplines / _compute_optimal_gcv_parameter

function

`scipy.interpolate._bsplines:_compute_optimal_gcv_parameter`

source: /scipy/interpolate/_bsplines.py:2182

Signature

def _compute_optimal_gcv_parameter(X, wE, y, w)

Summary

Returns an optimal regularization parameter from the GCV criteria [1].

Parameters

X : array, shape (5, n): 5 bands of the design matrix X stored in LAPACK banded storage.
wE : array, shape (5, n): 5 bands of the penalty matrix $W^{- 1} E$ stored in LAPACK banded storage.
y : array, shape (n,): Ordinates.
w : array, shape (n,): Vector of weights.

Returns

lam : float: An optimal from the GCV criteria point of view regularization parameter.

Notes

No checks are performed.

Aliases

scipy.interpolate._bsplines._compute_optimal_gcv_parameter