Scribd
Upload
3 views19 pages

Recap: 03 June 2023 10:22

The document discusses Singular Value Decomposition (SVD), a matrix factorization method that decomposes a matrix into three matrices: U, Σ, and V. It outlines various applications of SVD in fields such as machine learning, natural language processing, computer vision, and bioinformatics. Additionally, the document includes details on the geometric intuition behind SVD and methods for calculating it.

Uploaded by

u21ec159
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 views19 pages

Recap: 03 June 2023 10:22

The document discusses Singular Value Decomposition (SVD), a matrix factorization method that decomposes a matrix into three matrices: U, Σ, and V. It outlines various applications of SVD in fields such as machine learning, natural language processing, computer vision, and bioinformatics. Additionally, the document includes details on the geometric intuition behind SVD and methods for calculating it.

Uploaded by

u21ec159
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/ 19

Recap

03 June 2023 10:22

Session on SVD Page 1


Non-square Matrix
03 June 2023 10:23

Session on SVD Page 2


Session on SVD Page 3
Rectangular Diagonal Matrix
05 June 2023 14:05

A matrix that would be diagonal if it were square, but instead is rectangular due to extra rows
or columns of zeros.

Session on SVD Page 4


Session on SVD Page 5
What is SVD
03 June 2023 10:23

SVD is a matrix decomposition/factorization method that decomposes a matrix into three


other matrices. Given a matrix A, the singular value decomposition of A is usually written as:

A = UΣVT
Here:

• U and V are orthogonal matrices. U is the left singular vectors and V is the right singular
vectors.
• Σ is a diagonal matrix containing what we call the singular values.

Session on SVD Page 6


Applications of SVD
05 June 2023 05:56

1. Machine Learning and Data Science: SVD is used in Principal Component Analysis (PCA), a
technique for dimensionality reduction. This is helpful when dealing with high-dimensional
data. It's also used in various recommendation systems, for example in collaborative
filtering which is used in Netflix movie recommendation.

2. Natural Language Processing (NLP): SVD is used in Latent Semantic Analysis (LSA), a
technique for extracting the underlying meaning (semantic information) from textual data.
LSA uses SVD to reduce the dimensionality of a term-document matrix, which helps identify
relationships between terms and documents.

3. Computer Vision: In computer vision, SVD is used in image compression. By keeping only
the largest singular values and corresponding singular vectors, we can represent an image
using less data without losing too much information.

4. Signal Processing: SVD is used to separate useful signals from noise. This is useful in
applications like mobile communications and audio signal processing.

5. Numerical Linear Algebra: SVD is used for matrix inversion and solving systems of linear
equations. It is often a numerically stable way to solve ill-conditioned systems.

6. Psychometrics: In psychology and education, SVD is used in the construction and scoring of
psychological and educational tests, where it is often important to extract underlying latent
traits.

7. Bioinformatics: SVD and related techniques are often used to analyze gene expression data,
where it is important to identify the underlying patterns of gene activity.

8. Quantum Computing: SVD is also used in quantum state tomography to understand the
state of a quantum system.

Session on SVD Page 7


SVD The Equation
03 June 2023 10:23

Session on SVD Page 8


Relationship with Eigen Decomposition
03 June 2023 10:39

Session on SVD Page 9


Session on SVD Page 10
Session on SVD Page 11
Geometric Intuition
03 June 2023 10:23

Session on SVD Page 12


Session on SVD Page 13
Demo 1 [2x3]
05 June 2023 14:59

Session on SVD Page 14


Session on SVD Page 15
Demo 2 [3x2]
05 June 2023 15:50

Session on SVD Page 16


How to Calculate SVD
05 June 2023 08:03

Session on SVD Page 17


SVD in PCA
03 June 2023 10:23

Session on SVD Page 18


Session on SVD Page 19

You might also like