spectrochempy.NNMF

class NNMF(**kwargs)[source]

Performs a Non Negative Matrix Factorization of a NDDataset .

Parameters

  • dataset (|NDDataset|) – The data to be analysed.

  • Ci (|NDDataset|) – Initial concentration profile.

  • Sti (|NDDataset|) – Initial Spectral profile.

  • **kwargs – Optional keyword parameters. See Other Parameters below.

Returns

C, St (|NDDataset|) – Optimized concentration and spectral profile.

Other Parameters

  • tol (float, optional) – Tolerance for a relative stopping condition.

  • maxtime (float, optional) – Time limit.

  • maxiter (float) – Limit number of iterations.

  • verbose – Print calculation details

Notes

Algorithm based on

C.-J. Lin. Projected gradient methods for non-negative matrix factorization. Neural Computation, 19(2007), 2756-2779.

If you find this tool useful, please cite the above work. Author : Chih-Jen Lin, National Taiwan University Copyright (c) 2005-2008 Chih-Jen Lin See LICENCES in the root directory.

The current code is based on the Python translation by Anthony Di Franco: https://www.csie.ntu.edu.tw/~cjlin/nmf/others/nmf.py

Methods

NNMF.nlssubprob(V, W, Hinit, tol, maxiter)

Parameters

  • V, W -- Constant matrices.

NNMF.nmf(V, Winit, Hinit, tol, maxtime, maxiter)

NMF by alternative non-negative least squares using projected gradients.