Geometric distance and mean for positive semi-definite matrices of fixed rank

作者: Rodolphe Sepulchre , Silvère Bonnabel

DOI:

关键词:

摘要: This paper introduces a new metric and mean on the set of positive semidefinite matrices fixed-rank. The proposed is derived from well-chosen Riemannian quotient geometry that generalizes reductive cone associated natural metric. resulting space has strong geometrical properties: it geodesically complete, invariant with respect to all transformations preserve angles (orthogonal transformations, scalings, pseudoinversion). A meaningful approximation distance proposed, can be efficiently numerically computed via simple algorithm based SVD. induced preserves rank, possesses most desirable characteristics geometric mean, easy compute.

参考文章(18)
Jacques Faraut, Adam Korányi, Analysis on Symmetric Cones ,(1995)
Bernhard Schölkopf, Koji Tsuda, Jean-Philippe Vert, Kernel Methods in Computational Biology MIT Press. ,(2004)
R. Sepulchre, R. Mahony, P.-A. Absil, Optimization Algorithms on Matrix Manifolds ,(2007)
S.T. Smith, Covariance, subspace, and intrinsic Crame/spl acute/r-Rao bounds IEEE Transactions on Signal Processing. ,vol. 53, pp. 1610- 1630 ,(2005) , 10.1109/TSP.2005.845428
Xavier Pennec, Pierre Fillard, Nicholas Ayache, A Riemannian Framework for Tensor Computing International Journal of Computer Vision. ,vol. 66, pp. 41- 66 ,(2006) , 10.1007/S11263-005-3222-Z
Dénes Petz, Róbert Temesi, Means of Positive Numbers and Matrices SIAM Journal on Matrix Analysis and Applications. ,vol. 27, pp. 712- 720 ,(2005) , 10.1137/050621906
Inderjit S. Dhillon, Joel A. Tropp, Matrix Nearness Problems with Bregman Divergences SIAM Journal on Matrix Analysis and Applications. ,vol. 29, pp. 1120- 1146 ,(2007) , 10.1137/060649021
Maher Moakher, A Differential Geometric Approach to the Geometric Mean of Symmetric Positive-Definite Matrices SIAM Journal on Matrix Analysis and Applications. ,vol. 26, pp. 735- 747 ,(2005) , 10.1137/S0895479803436937