1: \begin{abstract}

2: The spectral decomposition of a real skew-symmetric matrix $A$

3: can be mathematically transformed into a specific structured

4: singular value decomposition (SVD) of $A$.

5: Based on such equivalence, a skew-symmetric Lanczos

6: bidiagonalization (SSLBD) method is proposed for the specific

7: SVD problem that computes extreme singular values and the

8: corresponding singular vectors of $A$, from which the eigenpairs

9: of $A$ corresponding to the extreme conjugate eigenvalues in

10: magnitude are recovered pairwise in real arithmetic.

11: A number of convergence results on the method are established, and

12: accuracy estimates for approximate singular triplets are given.

13: In finite precision arithmetic, it is proven that the

14: semi-orthogonality of each set of basis vectors and the

15: semi-biorthogonality of two sets of basis vectors

16: suffice to compute the singular values accurately.

17: A commonly used efficient partial reorthogonalization strategy

18: is adapted to maintaining the needed semi-orthogonality and

19: semi-biorthogonality.

20: For a practical purpose, an implicitly restarted SSLBD

21: algorithm is developed with partial reorthogonalization.

22: Numerical experiments illustrate the effectiveness and

23: overall efficiency of the algorithm.

24: \end{abstract}

25: