The Column Space and the Null Space of a Matrix

Suppose that A is a m n matrix. Then

dim Null(A) + dim Col(A) = n.

Why:

dim Null(A) = number of free variables in

row reduced form of A.
a basis for Col(A) is given by the columns
corresponding to the leading 1s in the row
reduced form of A.

The dimension of the Null Space of a matrix is

called the nullity of the matrix.
The dimension of the Column Space of a matrix is
called the rank of the matrix.

0-0
Linear transformations and matrices
Suppose that V and W are two vector spaces.
A linear transformation is a function f : V W
such that

f (rx + sy) = rf (x) + sf (y),

for all x, y V and r, s R.

Let A be an n m matrix.
Then A determines a linear transformation
fA : Rm Rn which is given by:
fA (x) = Ax, for x Rm .
To check this is is enough to observe that if x, y Rm
and r, s R then

fA (rx+sy) = A(rx+sy) = rAx+sAy = rfA (x)+sfA (y),

by the usual rules for matrix multiplication.

Hence, fA is a linear transformation as claimed.

0-1
Linear transformations: R R
Question :
When is a map f : R R a linear transformation?
Write f (1) = a, for some a R.
Then f (2) = f (1+1) = f (1)+f (1) = a +a = 2a
Alternatively, f (2) = f (2 1) = 2f (1) = 2a.
More generally, f (u) = uf (1) = ua, for any r R.
Thus, the linear transformations from R to R are the
maps f (u) = ua, for some a R.
Question : What does this correspond to geometrically?

0-2
Linear transformations: R R2
Question :
When is a map f : R R2 a linear transformation?
 
As R2 = { xy : x, y R } if f : R R2 is a map then
 
we can write f (1) = ab , for some a, b R.

As f is a linear transformation  
f (u) = f (u 1) = uf (1) = u ab
    
= ua bu = a
b u .

Notice that the last equality expresses f in terms of

matrix multiplication.

0-3
Linear transformations R2 R
Question :
When is a map f : R2 R a linear transformation?
n   
Recall that 1 , 0 } is a basis of R2 .
0 1

   
Write f 1 = a and f 0 =c
0 1

     
Then, f uv = uf 1
0 +vf 0
1 = ua+vc = (ua+vc)1
  
= a c uv .

Again, the last equality expresses f in terms of matrix

multiplication.

0-4
Linear transformations R2 R2
Question :
What are the linear transformations from R2 to R2 ?
       
As before, write f 10 = ab and f 01 = dc , for
some a, c R.
         
Then f uv = uf 10 + vf 01 = u ab + v dc
      
= uaub + vc
vd = a c
b d
u
v .

Once again, the last equality expresses f in terms of

matrix multiplication.
This shows that every linear map from R2 to R2 is
given by left multiplication by a 2 2 matrix.

0-5
 Linear transformations
Lets summarise the last four examples:
Linear transformations Description Matrix
    
RR f (u) = au u 7 a u

      
R R2 f (u) = u ab u 7 a b u

      
R2 R f uv = au + bv u
v
7 a b
u
v
          
R2 R2 f uv = u ab + v dc u a c
v 7 b d
u
v

In each of the cases above, the linear transformations

Rm Rn above are given by n m matrices, for
n, m {1, 2}.

This is true in general.

0-6
Linear transformations from Rm to Rn
Let f : Rm Rn be a linear transformation.
"1# " a11 # "0# " a21 #
0 a12 1 a22
Write f .. = .. , f .. = .. ,. . . ,
. . . .
0 a1n 0 a2n
" 0 # " am1 #
0 am2
f .. = .. .
. .
1 amn
" u1 # "1# "0# "0#
u2 0 1 0
Then f .. = u1 f .. +u2 f .. + +um f ..
. . . .
um 0 0 1
" a11 # " a21 # " am1 #
a12 a22 am2
= u1 .. +u2 .. + +um .. .
. . .
" a11a1n
a21 ...
a2n
am1 #" u1 #
amn
a12 a22 ... am2 u2
= .. .. . . .. .. .
. . . . .
a1n a2n ... amn um

That is, if x Rm then f (x) = fA (x), where

" a11 a21 ... am1 #
a12 a22 ... am2
A= .. .. . . .. .
. . . .
a1n a2n ... amn

0-7
Some properties of linear transformations
Let f : V W be a linear transformation.
Then
f (0) = 0.

f (x) = f (x), for all x V .

Remark: Note that the first statement says that

f (0V ) = 0W . That is, f has to map the zero vector of V
to the zero vector of W .
Why?
f (0) = f (0 0) = 0f (0) = 0.

f (x) = f (1 x) = f (x).

0-8
More properties of a linear transformation
Suppose that f : V W is a linear transformation and
set
X = { f (x) : x V } W.

Claim X is a vector subspace of W .

(1) X 6= since 0 = f (0) X.
(2) If x , y X then x = f (x) and y = f (y).
So that x + y = f (x) + f (y) = f (x + y) X.
(3) If x X and r R then x = f (x) for some
x V.
So rx = rf (x) = f (rx) X.
So X is a non-empty subset of W which is closed
under vector addition and scalar multiplication.
Hence, X is a vector subspace of W .

0-9
More examples of linear transformations
D is the vector space of differentiable functions.
Claim : Differentiation is a linear transformation on D.
This is the map D : D D give by D(f ) = f .
d d d
Check: D(rf + sg) = dx (rf + sg) = r dx f + s dx g
= rf + sg That is, D(rf + sg) = rD(f ) + sD(g).

Similarly, integration defines a linear transformation:

Rb
Iab (f ) = a f (x) dx.
We have, Rb
Iab (rf + sg) = a (rf (x) + sg(x)) dx
Rb Rb
= r a f (x) dx + s a g(x)) dx
= rIab (f ) + sIab (g).
Hence, Iab is a linear transformation.

0-10