MIT18.06 Linear Algebra (17)

lec17

Orthonormal vectors

qiqj={0 if ij1 if i=jq_{i}^\top q_j=\left\{\begin{array}{ll}0 & \text { if } i \neq j \\1 & \text { if } i=j\end{array}\right.

So Q=Q1Q^\top = Q^{-1}

We call QQ orthonormal matrix when it’s square.

Hadamard matrix

Let HH be the n order Hadamard matrix, then

is 2n order Hadamard matrix.

e.g.

H1=[1]H2=[1111]H4=[1111111111111111]\begin{align} H_1 & = \begin{bmatrix}1\end{bmatrix} \\ H_2 & = \begin{bmatrix} 1 & 1 \\ 1 & -1 \end{bmatrix} \\ H_4 & = \begin{bmatrix} 1 & 1 & 1 & 1 \\ 1 & -1 & 1 & -1 \\ 1 & 1 & -1 & -1 \\ 1 & -1 & -1 & 1 \end{bmatrix} \\ \vdots \end{align}

12H4\frac{1}{2}H_4 is an orthonormal matrix.

Gram-Schmidt

From any two vectors a,ba ,b , get two orthogonal vectors A,BA,B , and then get two orthonormal vectors q1,q2q_1, q_2 .

Gram

A=aA=a

B=bAbAAAB=b-\frac{A^\top b}{A^\top A}A

i.e. BB is the error vector ee in lec15, which is perpendicular to AA , i.e. AB=0A^\top B=0

C=cACAAABCBBBC = c - \frac{A^\top C}{A^\top A}A - \frac{B^\top C}{B^\top B}B

cc is a known vector.

Schmidt

q1:=AAq2:=BBq3:=CCq_1 := \frac{A}{\| A\| }\\ q_2 := \frac{B}{\| B\| }\\ q_3 := \frac{C}{\| C\| }\\

A=QRA=QR

A:=[ ab ]A:=\begin{bmatrix} \vdots & \vdots\\a & b\\ \vdots & \vdots\end{bmatrix}

Q:=[ q1q2 ]Q:=\begin{bmatrix} \vdots & \vdots\\q_1 & q_2\\ \vdots & \vdots\end{bmatrix}

then

where RR is an upper triangular matrix.

Because:

[a b]=[q1 q2][aq1bq1aq2bq2]\begin{bmatrix}a  & b\end{bmatrix} =\begin{bmatrix}q_1  & q_2\end{bmatrix}\begin{bmatrix}a^\top q_1 & b^\top q_1 \\a^\top q_2 & b^\top q_2\end{bmatrix}


MIT18.06 Linear Algebra (17)
https://yzzzf.xyz/2022/09/27/MIT16.08-lec17/
Author
Zifan Ying
Posted on
September 27, 2022
Licensed under