馃攳

Understanding Eigenvectors and Eigenvalues

Aug 8, 2024

Lecture on Eigenvectors and Eigenvalues

Introduction

  • Eigenvectors and eigenvalues are often found unintuitive by students.
  • The topic requires a solid understanding of prior concepts: matrices as linear transformations, determinants, linear systems of equations, and change of basis.
  • Proper visualization is crucial for understanding.

Linear Transformations in Two Dimensions

  • Example: A linear transformation moves basis vectors i-hat and j-hat to new coordinates, represented by a matrix.
    • i-hat to (3, 0)
    • j-hat to (1, 2)
    • Matrix representation: columns are [3, 0] and [1, 2]
  • Vectors usually get knocked off their span during transformation.
  • Special vectors stay on their span and are stretched or squished.
    • i-hat is a special vector stretched by a factor of 3.
    • Vector (-1, 1) is stretched by a factor of 2.

Eigenvectors and Eigenvalues

  • Special vectors that remain on their span are called eigenvectors.
  • The factor by which they are stretched or squished is the eigenvalue.
  • Example: Rotation in 3D
    • Eigenvector remains on its span, indicating the axis of rotation with eigenvalue 1.
  • Eigenvectors and eigenvalues give insight into the transformation beyond matrix representation.

Symbolic Representation

  • Expression: A * v = 位 * v, where
    • A is the matrix
    • v is the eigenvector
    • 位 is the eigenvalue
  • Rewriting the equation:
    • (A - 位I) * v = 0, where I is the identity matrix
    • Determinant of (A - 位I) must be zero for non-zero solutions.*

Computational Example

  • Example Matrix: columns [2, 1] and [2, 3]
  • Finding eigenvalues involves setting determinant to zero.
    • Subtract 位 from diagonal elements and compute determinant.
    • Example: 位 = 1 makes determinant zero.
  • Eigenvectors: solve (A - 位I) * v = 0
    • Example: Matrix [3, 0] and [1, 2]
    • Eigenvalues: 位 = 2 and 位 = 3
    • Eigenvectors: e.g., (-1, 1) for 位 = 2, stretched by factor of 2.*

Special Cases

  • Transformations without eigenvectors (e.g., 90-degree rotation): No real eigenvalues.
  • Shear transformation: Has eigenvalue 1, only vectors on x-axis are eigenvectors.
  • Diagonal matrices: Eigenvalues on diagonal, easier for computations.

Eigenbasis

  • Basis vectors as eigenvectors form an eigenbasis.
  • Diagonal matrices in an eigenbasis simplify power computations.
  • Change of basis: Use eigenvectors to form new basis coordinates.
    • Sandwich transformation matrix with change of basis matrix.
    • Resulting matrix is diagonal in the new coordinate system.

Conclusion

  • Eigenbasis makes matrix operations more manageable.
  • Not all transformations have enough eigenvectors for an eigenbasis.
  • Next lecture: abstract vector spaces.