tikhonov(U,s,V,b,lambda,x_0) Error

Question

Daniel Vasilaky 2013-7-5

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/81162-tikhonov-u-s-v-b-lambda-x_0-error

评论： ANA ROSHAN 2019-8-29

I cant get tikhonov(U,s,V,b,lambda,x_0) to run even for a simple example given in Christian Hansen Package. Here it is including the error:

EDU>> clear
EDU>> A= [
0.15  0.1
0.16  0.1
2.02  1.3
];
b =[
0.27
0.25
3.33];
x=[
1;1
];
b=A*x;
db=[.01; -.032; .01];
b1=b+db;
x2=pinv(A)*b1;
x_0=[0;0];
lambda=0.001;
[U,s,V]=svd(A);
[x_lambda,rho,eta]=tikhonov(U,s,V,b,lambda,x_0);
[x_lambda,rho,eta]=tikhonov(U,s,V,b1,lambda,x_0);
Index exceeds matrix dimensions.
Error in tikhonov (line 66)
  if (nargin==6), omega = V\x_0; omega = omega(1:p); end
%******************************************************************

When I try it on

A=hilb(12)
x=[1;1;1;1;1;1;1;1;1;1;1;1

after computong

b=A*x

and

U,s,V
x_0=[0;0;0;0;0;0;0;0;0;0;0;0]

I get NaN for output.

Not sure what I'm doing wrong. Can someone give a simple example that runs?

Thank you

2 个评论
显示无隐藏无

ANA ROSHAN 2019-8-29

VERY GOOD

ANA ROSHAN 2019-8-29

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Daniel Vasilaky 2013-7-11

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/81162-tikhonov-u-s-v-b-lambda-x_0-error#answer_91533

OK, I know what I was doing wrong. Instead of the full SVD I should have used the compact SVD, namely csvd. In a compact svd the columns of U and the rows of V’ are computed only for non-zero singular values. The zeroes in the full svd were causing the errors. Apologies for my ignorance.