fitting.nipals

fitting.nipals(A, n=-1)

Principal component analysis based on the NIPALS algorithm for deconvolution of spectra containing the signature of multiple species. The algorithm is based on Kubista et al., Crit. Rev. Anal. Chem. 29, 1 (1999) and references therein.

Parameters:

• A (array) – Data matrix with variable parameter in rows and wavelength / wavenumber in columns. This parameter can be concentration, temperature, time, etc.
• n (int) – Number of principal components to return. If n = -1, return all.

Returns:

The n most important target and projection vectors that are needed to reconstruct matrix A. Target vectors T are eigenvectors of A A’ with length being the eigenvalues. They are orthogonal. Projection vectors P are orthonormal. np.dot(T, P) yields the approximation to the matrix A.

See also

getRdimer() and getR23mer().