oemMeasurementCovmatNormalization

Workspace.oemMeasurementCovmatNormalization(self, oem: OptimalEstimationData = self.oem) → None

Returns measurement noise standard deviations \(D_{ii}=\sqrt{S_{\epsilon,ii}}\).

Stores these scales in oem.measurement_vec_normalization for measurement-space methods. The scaled system matrix is \(\mathbf{D}^{-1}(\mathbf{J}\mathbf{S}_a\mathbf{J}^{\top} +\mathbf{S}_\epsilon)\mathbf{D}^{-1}\). This does not change the statistical objective and is not full whitening for correlated measurement errors.

Author: Richard Larsson

Parameters:

oem (~pyarts3.arts.OptimalEstimationData, optional) – Numerical problem and results for oemCalc() and oemCalcReduced(). Defaults to self.oem. [INOUT]