Pattern Recognition on Oriented Matroids: Decompositions of Topes, and Orthogonality Relations
Andrey O. Matveev

TL;DR
This paper explores decompositions of topes in simple oriented matroids, establishing unique minimal subsets related to symmetric cycles and revealing orthogonality relations under specific conditions.
Contribution
It introduces a novel framework for decomposing topes via symmetric cycles and uncovers orthogonality relations based on subset sizes and parity conditions.
Findings
Unique minimal subsets Q(T,R) exist for topes T in simple oriented matroids.
Orthogonality relations hold for certain decompositions with large subset sizes and opposite parity ground sets.
The results connect to structural properties of tope graphs and symmetric cycles in oriented matroids.
Abstract
If V(R) is the vertex set of a symmetric cycle R in the tope graph of a simple oriented matroid M, then for any tope T of M there exists a unique inclusion-minimal subset Q(T,R) of V(R) such that T is the sum of the topes of Q(T,R). If for decompositions Q(T',R') and Q(T",R") with respect to symmetric cycles R' and R" in the tope graphs of two simple oriented matroids, whose ground sets have the cardinalities of opposite parity, we have |Q(T',R')|>3 and |Q(T",R")|>3, then these decompositions satisfy a certain orthogonality relation.
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsAdvanced Algebra and Logic · Advanced Graph Theory Research
Pattern Recognition on Oriented Matroids:
Decompositions of Topes,
and Orthogonality Relations
Andrey O. Matveev
Abstract.
If is the vertex set of a symmetric cycle in the tope graph of a simple oriented matroid , then for any tope of there exists a unique inclusion-minimal subset of such that is the sum of the topes of .
If for decompositions and with respect to symmetric cycles and in the tope graphs of two simple oriented matroids, whose ground sets have the cardinalities of opposite parity, we have and , then these decompositions satisfy a certain orthogonality relation.
1. Introduction
Let and be simple (i.e., with no loops, parallel elements or antiparallel elements) oriented matroids on their ground sets and , with sets of covectors and , and with sets of topes and , respectively; see [1] on oriented matroids. We suppose that
[TABLE]
Let be a symmetric -cycle, with its vertex set , in the tope graph of , and let be a symmetric -cycle in the tope graph of .
Let and be any topes of the oriented matroids and such that for the unique inclusion-minimal subsets and with the properties
[TABLE]
Let
[TABLE]
be the abstract simplicial complex on the vertex set , with the facet family
[TABLE]
where is the separation set of the topes and ; see e.g. [5] on such complexes.
Associate with the complex its “long” -vector
[TABLE]
defined by
[TABLE]
throughout the paper, the components of all vectors, as well as the rows and columns of matrices are indexed starting with zero.
Define vectors and by
[TABLE]
Let denote the square backward identity matrix of order , whose th entry is the Kronecker delta . We denote by the square forward shift matrix of order , whose th entry is .
According to the argument given in [5], and by [2, Prop. 3.51(a)], the vector
[TABLE]
is the long -vector of the boundary complex of a -dimensional simplicial convex polytope with vertices.
Associate with the abstract simplicial complex
[TABLE]
on the vertex set , with the facet family
[TABLE]
its long -vector
[TABLE]
defined by , .
The vector
[TABLE]
is the long -vector of the boundary complex of an -dimensional simplicial convex polytope with vertices.
Define the th entry of a square matrix
[TABLE]
of order to be .
Recall that the standard -vectors and of the boundary complexes and of simplicial polytopes both satisfy the Dehn–Sommerville relations [6, Sect. 8.3]
[TABLE]
As a consequence, the “long” -vectors
[TABLE]
and
[TABLE]
of the complexes and satisfy the Dehn–Sommerville type relations
[TABLE]
see e.g. [4, Sect. 2.3].
2. Orthogonality relations for decompositions of topes with respect to symmetric cycles in the tope graphs
Since the maximal face of the simplex does not belong to the complexes and , by [4, Eq. (2.3) of Prop. 2.1] we have
[TABLE]
for the vector
[TABLE]
where means the standard scalar product.
For the positive integers , define abstract simplicial complexes with their long -vectors to be the boundary complexes of the simplices by
[TABLE]
In view of (1.7), the long -vector lies either in the linear span
[TABLE]
when is even, or in the linear span
[TABLE]
when is odd; see [4, Sect. 3.1].
The Dehn–Sommerville type relations (1.6) and (1.7) imply that is a left eigenvector of the backward identity matrix that corresponds to its eigenvalue , while is a right eigenvector of that corresponds to the other eigenvalue . By the principle of biorthogonality [3, Th. 1.4.7(a)] we have
[TABLE]
In other words, together with the relation (2.3), the definitions (1.4) and (1.5) yield
[TABLE]
Note that the th entry of the square matrix
[TABLE]
of order is .
Let us sum up our conclusions:
Proposition 2.1**.**
Let and be simple oriented matroids on their ground sets and such that
[TABLE]
with sets of covectors and , and with sets of topes and , respectively.
Let be a symmetric cycle in the tope graph of , and let be a symmetric cycle in the tope graph of .
Let and be any topes of the oriented matroids and such that for the unique inclusion-minimal subsets and with the properties
[TABLE]
The long -vectors and of the complexes and whose families of facets are defined by (1.2) and (1.3), respectively, satisfy the orthogonality relation
[TABLE]
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1[1] Björner A. , Las Vergnas M. , Sturmfels B. , White N. , Ziegler G.M. Oriented matroids. Second edition. Encyclopedia of Mathematics, 46. – Cambridge: Cambridge University Press, 1999.
- 2[2] Gainanov D.N. Graphs for pattern recognition. Infeasible systems of linear inequalities. – Berlin: De Gruyter, 2016.
- 3[3] Horn R.A. , Johnson C.R. Matrix analysis. Second edition. – Cambridge: Cambridge University Press, 2013.
- 4[4] Matveev A.O. Pattern recognition on oriented matroids. – Berlin: De Gruyter, 2017.
- 5[5] Matveev A.O. Pattern recognition on oriented matroids: Decompositions of topes, and Dehn–Sommerville type relations. Preprint [ar Xiv:1703.04508], 2017.
- 6[6] Ziegler G.M. Lectures on polytopes. Revised edition. Graduate Texts in Mathematics, 152. – Berlin: Springer-Verlag, 1998.
