In mathematics, particularly linear algebra, an orthogonal basis for an inner product space ${\displaystyle V}$ is a basis for ${\displaystyle V}$ whose vectors are mutually orthogonal. If the vectors of an orthogonal basis are normalized, the resulting basis is an orthonormal basis.

## As coordinates

Any orthogonal basis can be used to define a system of orthogonal coordinates ${\displaystyle V.}$ Orthogonal (not necessarily orthonormal) bases are important due to their appearance from curvilinear orthogonal coordinates in Euclidean spaces, as well as in Riemannian and pseudo-Riemannian manifolds.

## In functional analysis

In functional analysis, an orthogonal basis is any basis obtained from an orthonormal basis (or Hilbert basis) using multiplication by nonzero scalars.

## Extensions

### Symmetric bilinear form

The concept of an orthogonal basis is applicable to a vector space ${\displaystyle V}$ (over any field) equipped with a symmetric bilinear form ${\displaystyle \langle \cdot ,\cdot \rangle }$, where orthogonality of two vectors ${\displaystyle v}$ and ${\displaystyle w}$ means ${\displaystyle \langle v,w\rangle =0}$. For an orthogonal basis ${\displaystyle \left\{e_{k}\right\))$:

${\displaystyle \langle e_{j},e_{k}\rangle ={\begin{cases}q(e_{k})&j=k\\0&j\neq k,\end{cases))}$
where ${\displaystyle q}$ is a quadratic form associated with ${\displaystyle \langle \cdot ,\cdot \rangle :}$ ${\displaystyle q(v)=\langle v,v\rangle }$ (in an inner product space, ${\displaystyle q(v)=\Vert v\Vert ^{2))$).

Hence for an orthogonal basis ${\displaystyle \left\{e_{k}\right\))$,

${\displaystyle \langle v,w\rangle =\sum _{k}q(e_{k})v^{k}w^{k},}$
where ${\displaystyle v_{k))$ and ${\displaystyle w_{k))$ are components of ${\displaystyle v}$ and ${\displaystyle w}$ in the basis.

The concept of orthogonality may be extended to a vector space over any field of characteristic not 2 equipped with a quadratic form ${\displaystyle q(v)}$. Starting from the observation that, when the characteristic of the underlying field is not 2, the associated symmetric bilinear form ${\displaystyle \langle v,w\rangle ={\tfrac {1}{2))(q(v+w)-q(v)-q(w))}$ allows vectors ${\displaystyle v}$ and ${\displaystyle w}$ to be defined as being orthogonal with respect to ${\displaystyle q}$ when ${\displaystyle q(v+w)-q(v)-q(w)=0}$.