In memory of Mohammed Bellalij
Abstract
The need to estimate upper and lower bounds for matrix functions of the form \({\mathrm{trace}}(W^Tf(A)V)\), where the matrix \(A\in {{\mathbb {R}}}^{n\times n}\) is large and sparse, \(V,W\in {{\mathbb {R}}}^{n\times s}\) are block vectors with \(1\le s\ll n\) columns, and f is a function arises in many applications, including network analysis and machine learning. This paper describes the shifted extended global symmetric and nonsymmetric Lanczos processes and how they can be applied to approximate the trace. These processes compute approximations in the union of Krylov subspaces determined by positive powers of A and negative powers of \(A-\sigma I_n\), where the shift \(\sigma\) is a user-chosen parameter. When A is nonsymmetric, transposes of these powers also are used. When A is symmetric and \(W=V\), we describe how error bounds or estimates of bounds for the trace can be computed by pairs of Gauss and Gauss-Radau quadrature rules, or by pairs of Gauss and anti-Gauss quadrature rules. These Gauss-type quadrature rules are defined by recursion coefficients for the shifted extended global Lanczos processes. Gauss and anti-Gauss quadrature rules also can be applied to give estimates of error bounds for the trace when A is nonsymmetric and \(W\ne V\). Applications to the computation of the Estrada index for networks and to the nuclear norm of a large matrix are presented. Computed examples show the shifted extended symmetric and nonsymmetric Lanczos processes to produce accurate approximations in fewer steps than the standard symmetric and nonsymmetric global Lanczos processes, respectively.
Similar content being viewed by others
Notes
If the matrix \(A\in {{\mathbb {R}}}^{n\times n}\) is symmetric the operator \(\backslash\) first seeks to compute the Cholesky factorization of \(A-\sigma I_n\). If this is not possible, because \(A-\sigma I_n\) is not positive definite, then an LU factorization is determined by Gaussian elimination with partial pivoting. The computed factorization is used to solve the linear system of equations with the matrix \(A-\sigma I_n\).
References
Alqahtani, H., Reichel, L.: Simplified anti-Gauss quadrature rules with applications in linear algebra. Numer. Algorithms 77, 577–602 (2018)
Alqahtani, H., Reichel, L.: Generalized block anti-Gauss quadrature rules. Numer. Math. 143, 605–648 (2019)
Baglama, J., Calvetti, D., Reichel, L.: IRBL: An implicitly restarted block-Lanczos method for large-scale Hermitian eigenproblems. SIAM J. Sci. Comput. 24, 1650–1677 (2003)
Baglama, J., Calvetti, D., Reichel, L.: Algorithm 827: irbleigs: A MATLAB program for computing a few eigenpairs of a large sparse Hermitian matrix. ACM Trans. Math. Softw. 29, 337–348 (2003)
Bai, Z., Golub, G.: Bounds for the trace of the inverse and the determinant of symmetric positive definite matrices. Ann. Numer. Math. 4, 29–38 (1997)
Baroni, S., Gebauer, R., Malcioglu, O.B., Saad, Y., Umari, P., Xian, J.: Harnessing molecular excited states with Lanczos chains. J. Phys. Condens. Mat., 22, Art. 074204 (2010)
Bellalij, M., Reichel, L., Rodriguez, G., Sadok, H.: Bounding matrix functionals via partial global block Lanczos decomposition. Appl. Numer. Math. 94, 127–139 (2015)
Bentbib, A.H., El Guide, M., Jbilou, K.: The block Lanczos algorithm for linear ill-posed problems. Calcolo 54, 711–732 (2017)
Bentbib, A.H., El Ghomari, M., Jagels, C., Jbilou, K., Reichel, L.: The extended global Lanczos method for matrix function approximation. Electron. Trans. Numer. Anal. 50, 144–163 (2018)
Bouyouli, R., Jbilou, K., Sadaka, R., Sadok, H.: Convergence properties of some block Krylov subspace methods for multiple linear systems. J. Comput. Appl. Math. 196, 498–511 (2006)
Caldarelli, G.: Scale-Free Networks. Oxford University Press, Oxford (2007)
De la Cruz Cabrera, O., Matar, M., Reichel, L.: Analysis of directed networks via the matrix exponential. J. Comput. Appl. Math 355, 182–192 (2019)
Davis, T., Hu, Y.: The SuiteSparse Matrix Collection. https://sparse.tamu.edu
Druskin, V., Knizhnerman, L.: Extended Krylov subspace approximations of the matrix square root and related functions. SIAM J. Matrix Anal. Appl. 19, 755–771 (1998)
Estrada, E.: Characterization of the folding degree of proteins. Bioinformatics 18, 697–704 (2002)
Estrada, E.: The Structure of Complex Networks: Theory and Applications. Oxford University Press, Oxford (2011)
Estrada, E., Fox, M., Higham, D., Oppo, G.L. (eds.): Network Science. Complexity in Nature and Technology, Springer, New York (2010)
Estrada, E., Hatano, N.: Statistical-mechanical approach to subgraph centrality in complex networks. Chem. Phys. Lett. 439, 247–251 (2007)
Estrada, E., Higham, D.J.: Network properties revealed through matrix functions. SIAM Rev. 52, 696–714 (2010)
Estrada, E., Rodriguez-Velazquez, J.A.: Subgraph centrality in complex networks. Phys. Rev. E. 71, Art. 056103 (2005)
Fenu, C., Martin, D., Reichel, L., Rodriguez, G.: Block Gauss and anti-Gauss quadrature with application to networks. SIAM J. Matrix Anal. Appl. 34, 1655–1684 (2013)
Gautschi, W.: Orthogonal Polynomials: Computation and Approximation. Oxford University Press, Oxford (2004)
Golub, G.H., Meurant, G.: Matrices, Moments and Quadrature with Applications. Princeton University Press, Princeton (2010)
Golub, G.H., Van Loan, C.F.: Matrix Computations, 4th edn. Johns Hopkins University Press, Baltimore (2013)
Han, I., Malioutov, D., Shin, J.: Large-scale log-determinant computation through stochastic Chebyshev expansions. In: Bach, F., Blei, D. (Eds.) Proceedings of The 32nd International Conference on Machine Learning, Lille, France, 2015, JMLR Workshop and Conference Proceedings, 37, pp. 908–917 (2015)
Jagels, C., Jbilou, K., Reichel, L.: The extended global Lanczos method, Gauss-Radau quadrature, and matrix function approximation, J. Comput. Appl. Math. 381 (2021) Art. 113027
Jagels, C., Reichel, L.: The structure of matrices in rational Gauss quadrature. Math. Comput. 82, 2035–2060 (2013)
Jbilou, K., Messaoudi, A., Sadok, H.: Global FOM and GMRES algorithms for matrix equations. Appl. Numer. Math. 31, 49–63 (1999)
Jbilou, K., Sadok, H., Tinzefte, A.: Oblique projection methods for multiple linear systems. Electron. Trans. Num. Anal. 20, 119–138 (2005)
Laurie, D.P.: Anti-Gaussian quadrature formulas. Math. Comp. 65, 735–747 (1996)
Newman, M.E.J.: Networks: An Introduction. Oxford University Press, Oxford (2010)
Ngo, T.T., Bellalij, M., Saad, Y.: The trace ratio optimization problem. SIAM Rev. 54, 545–569 (2012)
Pozza, S., Pranić, M.S., Strakoš, Z.: The Lanczos algorithm and complex Gauss quadrature. Electron. Trans. Numer. Anal. 50, 1–19 (2018)
Saad, Y., Chelikowsky, J., Shontz, S.: Numerical methods for electronic structure calculations of materials. SIAM Rev. 52, 3–54 (2010)
Sorensen, D.C.: Numerical methods for large eigenvalue problems. Acta Numer. 11, 519–584 (2002)
Ubaru, S., Chen, J., Saad, Y.: Fast estimation of \(tr(f(A))\) via stochastic Lanczos quadrature. SIAM J. Matrix Anal. Appl. 38, 1075–1099 (2017)
Varga, R.S.: Geršgorin and His Circles. Springer, Berlin (2004)
Acknowledgements
The authors would like to thank Carl Jagels and Giuseppe Rodriguez for carefully reading the manuscript and for comments that lead to an improved presentation. Research by LR was supported in part by NSF Grant DMS-1720259.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Bentbib, A.H., El Ghomari, M., Jbilou, K. et al. Shifted extended global Lanczos processes for trace estimation with application to network analysis. Calcolo 58, 4 (2021). https://doi.org/10.1007/s10092-020-00395-1
Received:
Revised:
Accepted:
Published:
DOI: https://doi.org/10.1007/s10092-020-00395-1
Keywords
- Shifted extended Krylov subspace
- Block Lanczos process
- Extended global symmetric Lanczos process
- Extended global nonsymmetric Lanczos process
- Gauss quadrature
- Anti-Gauss quadrature
- Network analysis
- Estrada index
- Trace estimation
- Nuclear norm estimation