How to Find Eigenvalues and Eigenvectors | Linear Algebra

Added:

Numerical Methods and the Instability of the Characteristic Polynomial

While introductory linear algebra teaches the analytical method of finding eigenvalues using the characteristic equation, det(A - λI) = 0, this approach is computationally impractical and numerically unstable for larger matrices. In practice, finding the roots of a high-degree polynomial is highly sensitive to rounding errors (known as Wilkinson's phenomenon). Furthermore, Abel's impossibility theorem proves that no general algebraic formula exists to find the roots of polynomials of degree five or higher. Consequently, applied mathematics and computer science reject the characteristic equation in favor of iterative numerical methods, such as the QR algorithm, Power Iteration, or Jacobi eigenvalue algorithm. These methods approximate eigenvalues and eigenvectors directly through matrix decompositions and iterative transformations, bypassing the characteristic polynomial entirely. Understanding this computational perspective is crucial for solving real-world, large-scale problems in physics, engineering, and data science.

Understanding of basic matrix operations, specifically matrix-vector multiplication and the identity matrix.
Ability to calculate the determinant of 2x2 and 3x3 square matrices.
Proficiency in solving systems of linear equations and finding the null space (kernel) of a matrix via row reduction.
Familiarity with basic algebraic concepts, particularly finding the roots of polynomial equations.
Matrix diagonalization and learning how to compute high powers of matrices efficiently.
Applying eigenvalues and eigenvectors to solve systems of linear ordinary differential equations.
Exploring advanced matrix decompositions, such as the Spectral Theorem and Singular Value Decomposition (SVD).
Investigating practical applications like Principal Component Analysis (PCA) in data science and finding normal modes in physics.
1.2M views23.5Klikes17:09@ProfessorDaveExplainsOriginal Release: 2019-07-17

Eigenvalues and eigenvectors are fundamental concepts in linear algebra where an eigenvector x of a square matrix A satisfies Ax = λx, with λ being the corresponding eigenvalue; to find them, one solves the characteristic equation det(A - λI) = 0 to obtain eigenvalues, then substitutes each eigenvalue back into (A - λI)x = 0 and solves the resulting system to find the associated eigenvectors, noting that any scalar multiple of an eigenvector is also an eigenvector.