Scribd
Upload
3 views16 pages

Final Presentation

The document presents a group presentation on Singular Value Decomposition (SVD), detailing its method, applications, advantages, and disadvantages. SVD is a numerical technique that decomposes a matrix into simpler components, facilitating tasks like dimensionality reduction and data compression. It also discusses the efficiency of SVD in various implementations and its computational challenges.

Uploaded by

Sanial Hasan Rahul
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
3 views16 pages

Final Presentation

The document presents a group presentation on Singular Value Decomposition (SVD), detailing its method, applications, advantages, and disadvantages. SVD is a numerical technique that decomposes a matrix into simpler components, facilitating tasks like dimensionality reduction and data compression. It also discusses the efficiency of SVD in various implementations and its computational challenges.

Uploaded by

Sanial Hasan Rahul
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 16

Good Morning!

Contents
Most. Mahmuda Khatun,
Introduction & Discussion of SVD Method
Roll: 15

Rana Pervez,
Discussion of SVD Method
Roll: 78

Md. Shidul Islam,


Application of SVD Method
Roll: 69

G. M. Tazmul Hossain,
Matlab Code
Roll: 50

Sanial Hasan Rahul,


Advantages of SVD Method
Roll: 42

Khadija Akter Jui,


Disadvantages & Efficiencies of SVD Method
Roll: 21
WELCOME TO OUR PRESENTATION
Group: 06
Presentation Topic: Singular Value Decomposition
Introduction & Discussion of SVD
Method:
A numerical method called Singular Value Decomposition (SVD) that decomposes a matrix into
three simpler matrices. Although, it is not specifically an eigenvalue method, but it is closely
related.
If A is an m×n matrix, then ATA is an n×n symmetric matrix with nonnegative eigenvalues. The
singular values of an m×n matrix are the square roots of the eigenvalues of ATA, and the 2-norm of
a matrix is the largest singular value. The Singular Value Decomposition (SVD) factors A into a
product of two orthogonal matrices and a diagonal matrix of its singular values.

where ∑=diag (σ1, σ2, σ3………σr), and σ1≥σ2≥σ3≥…σr are the positive singular values of
A.
Introduction & Discussion of SVD
Method:
Introduction & Discussion of SVD
Method:
Introduction & Discussion of SVD
Method:
Matlab Code:
clear all
close all
!!A=[4 0;3 -5];
n=input('Enter the system size');
for i=1:n
for j=1:n
A(i,j)=input(['Enter the entrey of the coefficent matix A for the row' num2str(i)]);
end
end
[U,S,V]=svd(A);
A
U
S
V
R=U*S*V';
R
ADVANTAGES
Dimensionality Reduction: SVD helps reduce the number of features in a
dataset while preserving important information, making it useful for
applications like Principal Component Analysis (PCA).
Data Compression: By approximating a matrix with fewer singular values,
SVD can significantly reduce storage space for large datasets, especially
in image compression.
Noise Reduction: SVD can help filter out noise by allowing the
reconstruction of data using only the most significant singular values.
Disadvantages
•SVD can be costly to compute, particularly for large matrices.
•It may require a lot of resources to factorize the matrix into three parts for
the computation.
•The SVD algorithm is sensitive to data missing values.
•The left and right singular vectors are not unique, but the singular values
derived from SVD are.
•It can take a lot of memory to store the entire SVD matrices(U,V and σ).
Specially for large matrices.
• SVD assumes that the relationships between features are linear, which can
limit its ability to capture complex non-linear relationships.
EFFICIENCY
Performance: For moderate-sized matrices, SVD is efficient and
effective. However, for very large datasets, alternatives such as
randomized SVD or truncated SVD can be more efficient.
Implementation: Many libraries and software tools (like NumPy,
SciPy, and TensorFlow) provide optimized implementations of
SVD, making it easier to apply in practice.
Parallelization: SVD can benefit from parallel computing
techniques, improving performance on large datasets.
Thank You!

You might also like