Section 6.

2: Orthogonal Sets

Ben Lantz

MATH 2210 Fall 2023

Definition. A set of vectors S = {u1 , u2 , . . . , un }, is an orthogonal set if every pair of vectors in S is orthogonal, that
is ui · uj = 0 for all i 6= j.

82 3 2 3 2 39
>
> 3 1 1>
>
>
<6 7 6 7 6 7>
2=
6 7 6 7 6 7
Example 1. Show that the set S = 6 1 7 , 6 2 7 , 6 2 7 is an orthogonal set.
>
> 4 5 4 5 4 5>>
>
: 1 7
>
;
1 2

w̅ in its
ñ 137C 1 17121 i i 0

Will 3 E 1 C21 117 E 0

Uz.li 1 t 2 f2 1 E 0
Definition. A set of vectors B = {u1 , u2 , . . . , un }, is an orthogonal basis of a vector space V , if B is an orthogonal
set and B is a basis for V .

Note: Every orthogonal set is inherently linearly independent, so to show that an orthogonal set is an orthogonal
basis we need to show either that:

• The set is spans the vector space.

• The number of vectors in the set matches the dimension of the vector space.

82 3 2 3 2 39
>
> 3 1 1>
>
>
<6 7 6 7 6 7>
2=
6 7 6 7 6 7
Example 2. S = 6 1 7 , 6 2 7 , 6 2 7 is an orthogonal basis for R3 .
>
> 4 5 4 5 4 5>>
>
: 1 7
>
;
1 2

3 and there are 3 vectors in

Since dim IR
S
linearly independent orthogonal
S is an orthogonal basis for R
Example 3. The standard basis for any vector space is an orthogonal basis for that vector space.

i
Theorem 1. The coordinates of a vector x in an orthogonal basis B = {u1 , u2 , . . . , un } are

2 3
c1
6 7
6 7
6 c2 7
[x]B = 6 7 where ci = x · ui
6 .. 7 ui · ui
6 . 7
4 5
cn

Proof. Let x = c1 u1 + c2 u2 + · · · + cn un , and consider x · ui .

E Ui c ñ caltz t Chin ñi

C Kini CzñzÑ t ciui.it t cnn.tt

0 0 0

x̅.li Cititi
Note A similarprocess
can show that any
Ci i Eu
orthogonal set is L I
 2 x̅ 3 82 3 2
>
w̅ it
3 2 its39>
1
6 >
> 3 1 >
6 7 <6 7 6 7 6 7>2
=
6 7 6 7 6 7 6 7
Example 4. Find 6 1 7 where S = 6 1 7 , 6 2 7 , 6 27 .
4 5 >
> 4 5 4 5 4 5>>
>
: 7
>
;
8 1 1 2
S

a a

c 2

1611E 1162 181 E 33

386 2
664
E E 121521 EKE

1 11 41 4 1
11
 y·u
Definition. Let u, y 2 Rn . The projection of y onto u is ŷ = u.
u·u

y·u
Note: ŷ is a scalar multiple of u where the scalar is
u·u
.
So j a
If L = Span{u}, then we write ŷ = projL y (the projection of y onto L).

We can decompose y into two parts y = ŷ + z:

(1) ŷ is a vector in the span of u

(2) z = y ŷ is a vector orthogonal to u

Note: The length of y ŷ, given by ||y ŷ|| is the distance from y to L = Span{u}.
 2 3 2 3
7 4
Example 5. Let y = 4 5 and u = 4 5.
6 2
Find ŷ, the projection of y onto u, and then write y as a sum of two vectors, one in Span{u} and one orthogonal
to u.

g a I I prois my

5 J E
ESpan w̅ orthogonal to w̅

E J J

L
i
 Note the similarities of the following:
y·u
• The projection of y onto a vector u: ŷ = u
u·u
2 3
c1
6 7
6 7
6 c2 7
• The coordinates of y relative to an orthogonal basis, [y]B = 6 7 where ci = y · ui .
6 .. 7 ui · ui
6 . 7
4 5
cn

We can think about the coordinates of a vector y in an orthogonal basis {u1 , u2 , . . . , un } as the sum of the projections
of y onto the basis vectors.

yity

Note that y = projL1 y + projL2 y where L1 and L2 are the spaces spanned by u1 and u2 respectively.
Definition. An orthonormal set (or orthonormal basis) is a an orthogonal set (or basis) where each vector is a unit
vector (of length 1).

Any orthogonal set can be turned into an orthonormal set by normalizing each vector (scaling to a unit vector).
82 3 2 3 2 39
>
> 3 1 1 >
>
>
<6 7 6 7 6 2 7> =
6 7 6 7 6 7
Example 6. Construct an orthonormal basis from the orthogonal basis of R3 , S = 6 1 7 , 6 2 7,6 2 7 .
>
> 4 5 4 5 4 5>>
>
: 1 7
>
;
1 2

FEET
I
Hill w̅ it it is

11h11 TEE Te
it

Hist FE
is

orthonormal basis FTP

ff.f Ef.fIffd
Theorem 2. The columns of a matrix U are orthonormal if and only if U T U = I.

h i
T
Proof. Consider U U where U = u1 u2 u3 .
2 3 2 3
uT1 u1 uT1 u1 uT2 u1 uT3
6 7 h i 6 7
T 6 T 7 6 7
U U = 6 u2 7 · u1 u2 u3 = 6 u2 uT1 u2 uT2 u2 uT3 7
4 5 4 5
T
u3 u3 uT1 u3 uT2 u3 uT3

if and only if w̅ is a unitvector

UTUIT UT.li 1
if and only if Ñ and it are orthogonal
Ui Uj Ui Nj 0

Theorem 3. If U has orthonormal columns then

(a) ||U x||= ||x||

(b) (U x) · (U y) = x · y

(c) (U x) · (U y) = 0 if and only if x · y = 0.

Note: The matrix U represents a transformation that preserves length and orthogonality.