100 likes | 109 Views
ExcelR is provided Data Analytics Training for over 5 universities, namely IIM, ISB, BITS Pilani, UNIMAS and Woxen School of Business and 40 premier educational institutions. Excelr is the only organization that is partnered with UNIMAS to provide Data analytics certification with global validation. <br>All our trainers are Data Analytics experts with 15 Years of experience and they hail from IIT, ISB and IIM. Being one of the best data Analytics Institutes, we offer Data Analytics certification course with tailor made course curriculum to suit the professionals as well as freshers. <br>We provide 100% placements with more than 400 participants being successfully placed with top MNCs namely E&Y, Accenture, VMWare, Infosys, IBM etc
E N D
Singular Value Decomposition (SVD)
Matrix Inverse If, A B I , identity matrix × = A1 - Then, B = Identity matrix : 0 1 0 0 ... 1 0 0 0 ... . . . . . . ... . 0 0 ...1 0 0 0 1 0 ...
X= UΣVT U and V are Orthonormal matrices rxr diagonal matrix with r non-zero diagonal elements
VT X U ∑ 1 0 0 0 0 3 0 0 0 0 4 0 0 2 0 0 0 0 0 0 0 0 1 0 0 1 0 0 0 0 0 -1 1 0 0 0 4 0 0 3 0 0 2.24 0 0 0 0 0 0 0 0 1 0 0 0 0 1 0 0.45 0 0 0 0.89 0 0 0 1 0 0 = x x 0 0 0 OPTIONAL OPTIONAL OPTIONAL 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)
Applications of SVD in image processing – closest rank-k approximation for a matrix X u X i ∑ = 1 k T k v Σ = i i i – 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 Original size 4*5=20 bytes VT X U ∑ 1 0 0 0 0 3 0 0 0 0 4 0 0 2 0 0 0 0 0 0 0 0 1 0 0 1 0 0 0 0 0 -1 1 0 0 0 4 0 0 3 0 0 2.24 0 0 0 0 0 0 0 0 1 0 0 0 0 1 0 0.45 0 0 0 0.89 0 0 0 1 0 0 = x x 0 0 0 k=1 0 x 4 x 0 0 1 0 1 0 0 0 = 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4 0 0 0 Compressed size 4*1+1+1*5=10 bytes
k=2 0 0 x 0 1 0 0 1 0 4 0 x 0 3 0 1 0 0 0 = 0 0 1 0 0 0 0 0 0 0 0 0 3 0 0 0 0 0 0 0 0 4 0 0 0 Compressed size 4*2+2+2*5=20 bytes k=3 0 0 1x 0 1 0 0 0 0 1 0 0 4 0 0 3 0 0 2.24 0x 0 0 1 0 0 0 1 0 0 0 0 0.89 0= 0 1 0 0 0 0 0 0 4 0 3 0 0 0 2 0 0 0 0 0 0 Compressed size 4*3+3+3*5=30 bytes 0.45 0 k=4 0 0 1 0 0 1 0 0 0 0 0 -1 1 0 0 0 4 0 0 0 0 3 0 2.24 0 0 0 0 0 0 0 0 0 1 0 0 0 0 1 0 0 0 0 0 0.89 1 0= 0 1 0 0 0 0 0 0 4 0 3 0 0 0 2 0 0 0 0 0 0 Compressed size 4*4+4+4*5=40 bytes x x 0.45 0 0 0
The image compression example in http://journal.batard.info/post/2009/04/08/svd- fun-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