profile pic
⌘ '
raccourcis clavier

every multivariate continuous function f:[0,1]nRf: [0,1]^n \to \mathbb{R} can be represented as superposition of continuous single-variable functions

definition

if ff is a multivariate continuous function, then ff can be written as a finite composition of continuous single-variable functions with binary operation of addition. Formally:

f(x)=f(x1,,xn)=q=02nΦq ⁣(p=1nϕq,p(xp))\displaystyle f(\mathbf {x} )=f(x_{1},\ldots ,x_{n})=\sum _{q=0}^{2n}\Phi _{q}\!\left(\sum _{p=1}^{n}\phi_{q,p}(x_{p})\right)

where ϕq,p:[0,1]R\phi_{q,p} : [0,1] \to \mathbb{R} and Φq:RR\Phi_q: \mathbb{R} \to \mathbb{R}

machine learning