In linear algebra, a projection decomposes a vector into a combination of two: one that is parallel to some line in the plane, and one that is perpendicular. The easiest formula is below, for the projection of onto .

\text{proj}_\vec u(\vec x)=\bigg(\frac{\vec u \cdot \vec x}{\|\vec u\|^2}\bigg)\vec u

Theorems

  • Suppose is a subspace of and . The closest vector in to is given by . In other words, for any .

See also