Skip to main content

TableĀ 2 Some common realizations of matrix factorization framework parameters

From: Deep data analysis via physically constrained linear unmixing: universal framework, domain examples, and a community-wide platform

Parameters Some common realizations
Similarity functions Frobenius norm, KL-divergence
Transformation function Logistic function, rounding function
Regularization Sparsity, spatial
Hard constraints Non-negativity, orthogonality, sum to one
Weights Uniform weights
Preprocessing Mean centering, normalization, log transformation, FFT