Properties of Projection Matrices
What you'll learn: You'll discover the special algebraic properties that make projection matrices unique and powerful tools in linear algebra.
The Core Idea
A projection matrix P that projects vectors onto a subspace has two remarkable properties that completely characterize it:
- Symmetry: P = P^T (the matrix equals its transpose)
- Idempotence: P² = P (projecting twice is the same as projecting once)
These aren't just mathematical curiosities—they capture the essence of what projection means geometrically.
Why These Properties Make Sense
Idempotence is intuitive: once a vector is projected onto a subspace, it's already in that subspace. Projecting it again doesn't move it anywhere new—it stays put. Mathematically, if v is already in the subspace, then Pv = v, so P(Pv) = Pv.
Symmetry comes from the orthogonality condition. When we project onto a subspace using an orthonormal basis, the resulting matrix has equal components above and below its diagonal. This reflects the fact that the inner product used to compute projections is symmetric: ⟨u, v⟩ = ⟨v, u⟩.
Additional Properties
From these two properties, several others follow:
- Eigenvalues are only 0 or 1: Since P² = P, any eigenvalue λ satisfies λ² = λ, meaning λ(λ - 1) = 0
- Rank equals trace: The rank of P (dimension of the subspace) equals its trace (sum of diagonal entries)
- I - P is also a projection matrix: It projects onto the orthogonal complement
These properties let you verify whether a matrix is a valid projection matrix and understand its geometric behavior.
Key Takeaway: Orthogonal projection matrices are uniquely characterized by being both symmetric (P = P^T) and idempotent (P² = P), properties that directly reflect the geometric nature of projection.