Abstract
We introduce a new majorization order for classes (sets) of matrices which generalizes several existing notions of matrix majorization. Roughly, the notion says that every matrix in one class is majorized by some matrix in the other class. The motivation to study this majorization concept comes from mathematical statistics and involves the information content in experiments. This connection is briefly described. We investigate properties of this new order both of algebraic and geometric character. In particular, we establish results on so-called minimal cover classes with respect to the introduced majorization.
Original language | English |
---|---|
Pages (from-to) | 201-221 |
Number of pages | 21 |
Journal | Linear Algebra and Its Applications |
Volume | 555 |
DOIs | |
State | Published - 15 Oct 2018 |
Externally published | Yes |
Bibliographical note
Publisher Copyright:© 2018 Elsevier Inc.
Funding
The investigations of the second and the third authors were financially supported by RSF grant 16-11-10075.
Funders | Funder number |
---|---|
Russian Science Foundation | 16-11-10075 |
Keywords
- Doubly stochastic matrix
- Matrix majorization
- Partial order