Graphical Singular Value Decomposition SVD

singular value decomposition

Purpose

Decompose an arbitrary matrix M into the product form of three matrices
insert image description here

linear transformation

Stretch:
insert image description here

Rotation:

Left multiply an orthogonal matrix
insert image description here

Physical meaning of decomposition

insert image description here
insert image description here
insert image description here

generalizes to matrices of any size

insert image description here
The singular value decreases sequentially from top to bottom, and the part with small singular value can be removed for data compression.

The effect after compression:
insert image description here

Compressed reconstruction (Using U, V and Σ with some values ​​removed)

insert image description here
insert image description here

Calculation of SVD decomposition

insert image description here

Introduce the calculation of eigenvalues
insert image description here
insert image description here
insert image description here

Calculation process

insert image description here

Guess you like

Origin blog.csdn.net/gary101818/article/details/123665697