# data science training i delhi

Singular Value Decomposition (SVD)

Matrix Transpose

Matrix Multiplication

Matrix Inverse If, A× B = I ,identity matrix Then, B= A -1 Identity matrix : 1 0 ... 0 0 0 1 ... 0 0   . . . . . . ... .    0 0 ...1 0 0 0 ... 0 1  

X=

T UΣV

U and V are Orthonormal matrices

rxr diagonal matrix with r non-zero diagonal elements

1 0 0 0

0 0 0 4

X 0 3 0 0

0 0 0 0

2 0 0 0 = 0 0 0 1

U 0 1 0 0

∑ 1 0 0 0

0 4 0 0 0 0 0 0 3 0 0 0 x x -1 0 0 2.24 0 0.45 0 0 0 0 0 0

OPTIONAL R Code > M=matrix(c(1,0,0,0,0,0,0,4,0,3,0,0,0,0,0,0,2,0,0,0),nrow=4,ncol=5) > X=svd(M) > X\$u > X\$d > X\$v > X\$u%*%diag(X\$d)%*%t(X\$v)

T

V 1 0 0 1 0 0 0 0

0 0 0 0 0 0.89 1 0

Applications of SVD in image processing – closest rank-k approximation for a matrix X k

X

k

= ∑ u i Σi vi

T

i =1

– Each term in the summation expression above is called principal image

Original matrix (X) 1 0 0 0 2 0 0 3 0 0 0 0 0 0 0 0 4 0 0 0

1 0 0 0

0 0 0 4

X 0 3 0 0

k=1 0x 4x 0 0 1

0 0 0 0

2 0 0 0 = 0 0 0 1

Original size 4*5=20 bytes

U 0 1 0 0

T

∑ 1 0 0 0

0 4 0 0 0 0 0 0 3 0 0 0 x x -1 0 0 2.24 0 0.45 0 0 0 0 0 0

0 1 0 0 0=

0 0 0 0

0 0 0 4

0 0 0 0

0 0 0 0

0 0 0 0

V 1 0 0 1 0 0 0 0

0 0 0 0 0 0.89 1 0

Compressed size 4*1+1+1*5=10 bytes

k=2 0 0x 0 1 0 0 1 0 k=3 00 01 00 10

1x 0 0 0

k=4 00 01 00 10

1 0 0 0

40x 0 1 0 0 0= 0 0 03 0 0 1 0 0 0 0 0 0 0 4

40 0x 03 0 0 0 2.24

0 0 x -1 0

4 0 0 0

0 0 3 0 0 2.24 0 0

0 1 0 0 0.45 0

0 0 x 0 0

0 0 0.45 0

0 1 0

0 3 0 0

0 0= 0 0 0 0.89

1 0 0 0

0 1 0 0

0 0= 0 0 0 0.89 1 0

0 0 0 0

1 0 0 0

0 0 0 0

0 0 0 4

Compressed size 4*2+2+2*5=20 bytes

0 3 0 0

1 0 0 0

0 0 0 0

0 0 0 4

2 0 0 0

0 3 0 0

Compressed size 4*3+3+3*5=30 bytes

0 0 0 0

2 0 0 0

Compressed size 4*4+4+4*5=40 bytes

The image compression example in http://journal.batard.info/post/2009/04/08/svdfun-profit

• Original size = 384*384 bytes = 147,456 bytes • k=1: 384*1+1+1*384=769 bytes • k=10: 384*10+10+10*384=7,690 bytes • k=20: 384*20+20+20*384=15,380 bytes • k=50: 384*50+50+50*384=38,450 bytes • k=100: 384*100+100+100*384=76,900 bytes • k=200: 384*200+200+200*384=153,800 bytes

## data science training i delhi

Singular Value Decomposition (SVD) Matrix Transpose Matrix Multiplication Matrix Inverse If, A× B = I ,identity matrix Then, B= A -1 Identity ma...

#### Recommend Documents

Data Science Training in Delhi
Excelr is the best technology training Data Science Training certification providing all the resources for youreffectiv

Data Science Course in delhi
Best data science training Institute: ExcelR is the best data science Course in Delhi.we provide real time project train

data science training
Learn data science training with Industry Experts and Get Placed in Top coprporates with the placement assistance provid

data science training bangalore
Data for a Data Scientist is what Oxygen is to Human Beings. This is also a profession where statistical adroit works

data science training
Get data science training and certfication with Industry Experts and Get Placed in Top corporates s with the placement a

Data science training
ExcelR offers 160 hours classroom training on Business Analytics / Data Scientist / Data Analytics. We are considered as

data science training
If you are looking for data science in egypt. You are at the right place - ExcelR ExcelR is global training provider pe

data science training
Learn data science training with certification with Industry Experts and Get Placed in Top coprporates with the placemen

data science training
If you are looking for data science in jeddah saudi arabia. You are at the right place - ExcelR ExcelR is global traini