Characterization and construction of sign regular tridiagonal matrices (Q6939439)

This interesting paper studies characterizations and constructions of tridiagonal sign regular (SR) matrices. SR matrices have the distinct property that they are matrices whose minors of the same order have the same sign. It turns out that SR matrices have multiple applications in many scientific areas, for example economics, approximation theory, statistics and many others. A key fact about SR matrices is that when these matrices are non-singular, they are characterized as variation-diminishing linear maps: That is, the number of sign changes in the consecutive components of the image of a vector is bounded above by the number of sign changes in the consecutive components of the vector. Tridiagonal SR matrices as it turns out play an important role in computer aided geometric design. We refer the interested reader to the references in the paper under review for a detailed account of all the applications and others described above for SR and tridiagonal SR matrices.\N\NThe authors provide a method for the construction of tridiagonal SR matrices, as well as necessary conditions to carry out this construction. They also provide a characterization of tridiagonal SR matrices in terms of the last elements of the matrix.