An implicit Jacobi-like method for computing generalized hyperbolic SVD Original Research Article

In this paper, we introduce a joint hyperbolic-orthogonal decomposition of two matrices which we call a generalized hyperbolic singular value decomposition, or GHSVD. This decomposition can be used for finding the eigenvalues and eigenvectors of a symmetric definite pencil XTX−λYTΦY where Φ=diag(±1). We also present an implicit Jacobi-like method for computing this GHSVD.