For a given nonlinear mappingΦ(𝑥), the input data spaceℝ𝑛can be mapped into the Feature space
Φ : ℝ𝑛 ↦ Feature space 𝑥 ↦ Φ(𝑥)
The feature map is an arbitrary map into thefeature space𝐻.
For example, if𝐻isfinite dimensional(say, of dimension𝑚), you can pick anorthonormal basisfor it,
and think ofeach component ofΦin that basisas afeature. This is what scalar features are.
Inner product space
orthonormal basis
In mathematics, particularly linear algebra, an orthonormal basis for an inner product space V with finite dimension is a basis for V whose vectors are orthonormal, that is, they are all unit vectors and orthogonal to each other.
For example, the standard basis(표준기저) for a Euclidean space R" is an orthonormal basis, where the relevant inner product is the dot product of vectors. The image(공역) of the standard basis under a rotation or reflection (or any orthogonal transformation) is also orthonormal(=standard basis를 가지고 rotation과 같은 orthogonal transformation을 하면 그 transformed vector들도 orthonormal이다), and every orthonormal basis for R" arises in this fashion.
*standard basis: 유클리드 공간에서 직교 좌표계의 축을 향하는 단위 벡터의 집합
Every vector a in three dimensions is a linear combination of the standard basis vectors i, j and k.
* basis: 벡터 공간을 선형생성하는 선형독립인 벡터들이다. 달리 말해, 벡터 공간의 임의의 벡터에게 선형결합으로서 유일한 표현을 부여하는 벡터들
For a general inner product space V, an orthonormal basis can be used to define normalized orthogonal coordinates on V. Under these coordinates, the inner product becomes a dot product of vectors. (The inner product는 dot product의 일반화된 버전 toabstract vector spacesover afieldofscalars, being either the field ofreal numbers or the field ofcomplex numbers)
orthogonal (직교)
orthonormal
Two vectors are said to be orthogonalif their dot product is zero Orthogonal vectors areperpendicularto each other