| Review: |
This book was written for a First-Semester graduate course in matrix theory at North Carolina State University. It covers linear system solution, least squares problems and singular value decomposition in detail. The book differs from other numerical linear algebra textbooks in that: it offers a systematic development of numerical conditioning; it hardly mentions floating-point arithmetic; it uses a simplified concept of numerical stability to give quantitative intuition while avoiding tedious round-off error analysis; it uses simple derivations; it takes a high-level view of algorithms thus omitting detailed implementation of direct methods; it gives results for complex rather than real matrices; it includes exercises that contain many useful facts. There are chapters on: matrices; sensitivity, errors and norms; linear systems; singular value decomposition; least-squares problems; and subspaces. |