Orthogonality
View .md

Projections

The projection of b\mathbf{b} onto a\mathbf{a}:

p=a⋅ba⋅aa=aaTaTab\mathbf{p} = \frac{\mathbf{a} \cdot \mathbf{b}}{\mathbf{a} \cdot \mathbf{a}}\mathbf{a} = \frac{\mathbf{a}\mathbf{a}^T}{\mathbf{a}^T\mathbf{a}}\mathbf{b}

Projection Matrix

P=aaTaTaP = \frac{\mathbf{a}\mathbf{a}^T}{\mathbf{a}^T\mathbf{a}}

Properties: P2=PP^2 = P, PT=PP^T = P (idempotent and symmetric).

Projection onto a Subspace

For projecting onto C(A)C(A):

P=A(ATA)−1ATP = A(A^TA)^{-1}A^T

Key Property

The error e=b−p\mathbf{e} = \mathbf{b} - \mathbf{p} is orthogonal to the projection:

e⊥p\mathbf{e} \perp \mathbf{p}