Publication Date:
2021
abstract:
The common principal components (CPC) and the proportional principal components (PPC) models are two possible generalizations of the standard PCA for several covariance matrices. The goal of this chapter is to revisit the classical CPC and PPC estimation based on the ML principle, and compare their features to their LS counterparts. The original CPC and PPC models are designed to produce full set of components.
Iris type:
2.1 Contributo in volume (Capitolo o Saggio)
List of contributors:
Trendafilov, Nickolay; Gallo, Michele
Book title:
Multivariate Data Analysis on Matrix Manifolds (with Manopt)