Let \(n_ +(X)\), \(n_ -(X)\) and \(n_ 0(X)\) denote the number of positive, negative and zero eigenvalues, respectively, of the real symmetric matrix \(X\). Then \(\text{sgn} X=n_ +(X)-n_ -(X)\) is the signature of \(X\). In the paper an arbitrary invertible \((m+n)\times(m+n)\) real symmetric matrix \(M\) and its inverse \(M^{-1}\) are considered in the partitioned form \(M=\left[\begin{matrix} A &B \\ B^ \top &C\end{matrix}\right]\) and \(M^{-1}=\left[\begin{matrix} P &Q \\ Q^ \top &R \end{matrix}\right]\) where \(A\) and \(P\) are \(m\times m\), and \(C\) and \(R\) are \(n\times n\) symmetric matrices. It is shown that \(\text{sgn} M=\text{sgn} A=\text{sgn} R\).
- Inverses of 2 2 block matrices
- A restricted signature normal form for Hermitian matrices, quasi-spectral decompositions, and applications
- The inverse of a partitioned positive-definxte symmetric matrix
- Using the Matrix Sign Function to Compute Invariant Subspaces
- The symmetric signature
- Spectral aspects of symmetric matrix signings
- On the signature of an involution
- Symmetric sign pattern matrices that require unique inertia
- On the expected number of critical points of locally isotropic Gaussian random fields
This page was built for publication: Signature of invertible symmetric matrices
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1813152)
