Abstract:The singular values $\sigma >1$ of an $n \times n$ involutory matrix $A$ appear in pairs $(\sigma, \frac{1}{\sigma}),$ while the singular values $\sigma = 1$ may appear in pairs $(1,1)$ or by themselves. The left and right singular vectors of pairs of singular values are closely connected. This link is used to reformulate the singular value decomposition (SVD) of an involutory matrix as an eigendecomposition. This displays an interesting relation between the singular values of an involutory matrix and its eigenvalues. Similar observations hold for the SVD, the singular values and the coneigenvalues of (skew-)coninvolutory matrices.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is related to the singular value decomposition (SVD) characteristics of (skew -) involutory matrices and (skew -) co - involutory matrices. Specifically, the author explores the relationships between the singular values and singular vectors of these special types of matrices and reveals the interesting connections between them and the eigenvalues.
### Main problems include:
1. **Pairing phenomenon of singular values**:
- For an involutory matrix \( A \), its singular values \( \sigma>1 \) always appear in pairs of the form \( (\sigma,\frac{1}{\sigma}) \).
- The singular value \( \sigma = 1 \) may appear alone or in pairs, depending on the relationship between the left - singular vector and the right - singular vector.
2. **Relationships between singular vectors**:
- If \( (u, v,\sigma) \) is a singular triple, then \( (v, u,\frac{1}{\sigma}) \) is also a singular triple.
- For \( \sigma = 1 \), there may be cases such as \( (u, u, 1) \) or \( (u, -u, 1) \).
3. **Relationship between singular value decomposition and eigenvalue decomposition**:
- Through the structure of SVD, the SVD of an involutory matrix can be reformulated as an eigenvalue decomposition, thus showing the relationship between singular values and eigenvalues.
4. **Extension to other types of matrices**:
- The paper also studies the SVD characteristics of skew - involutory matrices, co - involutory matrices and skew - co - involutory matrices and shows results similar to those of involutory matrices.
### Formula representation
- An involutory matrix \( A \) satisfies \( A^{2}=I_{n}\), that is, \( A = A^{-1}\).
- Singular value decomposition formula: \( A = U\Sigma V^{H}\), where \( U \) and \( V \) are unitary matrices, and \( \Sigma \) is a diagonal matrix containing non - negative real singular values.
- For singular values \( \sigma>1 \), there exists a pairing relationship \( \sigma_{1}=\frac{1}{\sigma_{n}}, \sigma_{2}=\frac{1}{\sigma_{n - 1}},\ldots,\sigma_{n}=\frac{1}{\sigma_{1}}\).
- For \( \sigma = 1 \), there may be paired singular triples \( (u, v, 1) \) and \( (v, u, 1) \), and there may also be a single singular triple \( (u,\pm u, 1) \).
Through these analyses, the paper aims to provide a comprehensive understanding of the singular value decomposition of these four types of matrices and reveal their inherent mathematical structures.