A transformation (mapping) L:X→YL:X→Y from a vector space XX
to a vector space YY (with the same scalar field KK) is a linear transformation if:
- L(αx)=αL(x)L(αx)=αL(x)∀x∈X∀x∈X, α∈Kα∈K
- L(x1+x2)=L(x1)+L(x2)L(x1+x2)=L(x1)+L(x2)∀x1,x2∈X∀x1,x2∈X.
We call such transformations linear operators.
- X=RNX=RN, Y=RMY=RML:RN→RML:RN→RM is an M×NM×N matrix
- Fourier transform: F(x(t))=∫-∞∞x(t)e-jwtdtF(x(t))=∫-∞∞x(t)e-jwtdtF:L2(R)→L2(R)F:L2(R)→L2(R)
Let L:X→YL:X→Y be an operator (linear or otherwise). The range spaceR(L)R(L) is
R
(
L
)
=
{
L
(
x
)
∈
Y
:
x
∈
X
}
.
R
(
L
)
=
{
L
(
x
)
∈
Y
:
x
∈
X
}
.
(1)
The null spaceN(L)N(L), also known as “kernel”, is
N
(
L
)
=
{
x
∈
X
:
L
(
x
)
=
0
}
.
N
(
L
)
=
{
x
∈
X
:
L
(
x
)
=
0
}
.
(2)
If LL is linear, then both R(L)R(L) and N(L)N(L) are subspaces.