This is an internal function that is only exposed on the public API for unit testing purposes. It computes the log-likelihood of the spline and the noise, once the spectral signature has been subtracted from the observed data. Thus, it can be used with either Lorentzian, Gaussian, or pseudo-Voigt broadening functions.
computeLogLikelihood(obsi, lambda, prErrNu, prErrSS, basisMx, eigVal,
precMx, xTx, aMx, ruMx)
Vector of residuals after the spectral signature has been subtracted.
smoothing parameter of the penalised B-spline.
hyperparameter of the additive noise
hyperparameter of the additive noise
Matrix of B-spline basis functions
eigenvalues of the Demmler-Reinsch factorisation
precision matrix for the spline
sparse matrix cross-product
orthoganal matrix A from the Demmler-Reinsch factorisation
product of Ru from the Demmler-Reinsch factorisation
The logarithm of the likelihood.