Change of Basis for Linear Transformations
You'll learn: How the matrix representation of a linear transformation changes when you switch to a different basis.
The Core Idea
You've seen that a linear transformation can be represented as a matrix—but which matrix you get depends on your choice of basis. Just as a vector's coordinates change when you switch bases, so does the matrix of a linear transformation.
Here's the key relationship: if a transformation T has matrix A in the old basis and you change to a new basis using transition matrix P, then the matrix of T in the new basis is:
B = P⁻¹AP
This formula is called a similarity transformation. The matrices A and B represent the same transformation, just viewed through different coordinate systems.
Why This Matters
Think of it like describing a rotation: you can describe it using standard x-y coordinates, or you can tilt your coordinate system first. The rotation itself doesn't change—only your numerical description of it changes.
The process works like this:
- Start with coordinates in the new basis
- Convert to the old basis using P
- Apply the transformation using A
- Convert back to the new basis using P⁻¹
Together: P⁻¹ · A · P = B
This is why similar matrices (related by B = P⁻¹AP) share important properties like eigenvalues, determinant, and trace—they're just different views of the same transformation.
Key Takeaway: The matrix of a linear transformation changes with the basis according to B = P⁻¹AP, where P is the transition matrix. This similarity transformation preserves the transformation's essential properties while adapting its representation to a new coordinate system.