Improvement for Nonnegative PCA Algorithm for Independent Component Analysis
Yunxia Li, Hong Yu Zheng · 2006
This paper consider the independent component analysis problem, in the case where the hidden sources are nonnegative and well-grounded. It makes improvement for nonnegative PCA algorithm by adding a term to the cost function to assure the orthonormality of the separating matrix. Simulation results illustrate its effectiveness