Course contentsShow
← Mathematics
Lesson 2256 of 3,12147. Linear Algebra — Applications and Advanced TopicsFree lesson

Similarity Transformations

Understanding similar matrices and their relationship through change of basis.

Similarity Transformations

You'll learn how similar matrices arise from representing the same linear transformation in different coordinate systems.

What Is a Similarity Transformation?

A similarity transformation is the process that relates two matrices representing the same linear transformation but expressed in different bases. If matrix AA represents a transformation in one basis and matrix BB represents the same transformation in another basis, we say AA and BB are similar.

Mathematically, matrices AA and BB are similar if there exists an invertible matrix PP (a transition matrix) such that:

B=P−1APB = P^{-1}AP

Here, PP is the change-of-basis matrix between the two coordinate systems. This equation encapsulates the entire similarity transformation.

Why Does This Matter?

Think of similarity transformations as changing your perspective on the same object. Imagine describing the location of furniture in a room using coordinates measured from the north wall versus from the door—different numbers, same furniture, same room.

Similarly, when you change basis:

  • The transformation itself doesn't change
  • Only the matrix representation changes
  • Key properties (eigenvalues, determinant, trace, characteristic polynomial) remain invariant

This is why similar matrices share eigenvalues—they're really the same transformation viewed through different lenses!

The Connection to Diagonalization

Diagonalization is actually a special case of similarity transformation. When you diagonalize A=PDP−1A = PDP^{-1}, you're finding a basis (the eigenvector basis) where the transformation looks particularly simple—just scaling along coordinate axes.

Key Takeaway: Similarity transformations formalize the idea that the same linear transformation can be represented by different matrices in different bases, connected by B=P−1APB = P^{-1}AP, preserving all essential properties of the transformation.