CODEWITHSUNDEEP
pythonnumpylinear-algebraleast-squaressvd
user2906838
user2906838

Reputation: 1178

Determining the corresponding columns in V of singular values in Numpy SVD

I'm trying to use SVD to estimate the solution for non-square matrix of linear equations. My matrix is of the 8 x 6 shape.

I calculated the following parameters using: 

U, sigma, VT = np.linalg.svd(mat)

Now, I am suggested to take a column from V with the smallest the corresponding value in S, and that should be the solution to my 6 parameters that I'm trying to determine with 8 equations.

Can somebody please help?

Thanks

Upvotes: 0

Views: 655

Answers (1)

user2906838
user2906838

Reputation: 1178

Oh! sorry guys, Thank you for your time. Actually I sorted out the problem with the slightly different answer here about the least square estimation of over-determined condition.

And this seems to do, I simply had to do: 

sol_min = VT[:, np.argmin(sigma)]

Upvotes: 1

Related Questions