SLIDE 1 NLA Reading Group Spring ’13
by Onur Güngör
Lecture 10 Householder Triangularization
SLIDE 2
Householder: orthogonal triangularization
Householder and Gram-Schmidt
Gram-Schmidt: triangular orthogonalization
SLIDE 3
Triangularization by Introducing Zeros
SLIDE 4
Householder Reflectors
SLIDE 5
Householder Reflectors
P is the projector onto the space H
SLIDE 6
Householder Reflectors
Instead of We use for numerical stability.
SLIDE 7
Householder Algorithm
SLIDE 8
Applying Q
This will be employed while solving least squares problems using QR factorization.
SLIDE 9
Forming Q
Q can be formed by calculating Qe1, Qe2, … and Qem.
SLIDE 10
Operation Count
Let Each vector requires flops.
SLIDE 11
Operation Count
SLIDE 12
Operation Count
SLIDE 13 NLA Reading Group Spring ’13
by Onur Güngör
Lecture 11 Least Squares Problems
SLIDE 14
Definition
SLIDE 15
Polynomial Interpolation
SLIDE 16 Polynomial Least Squares Fitting
Solve by minimizing
SLIDE 17
Orthogonal Projection
SLIDE 18
Pseudoinverse and Normal Equations
SLIDE 19
Least Squares via Normal Equations
SLIDE 20
Least Squares via QR Factorization
SLIDE 21
Least Squares via SVD