Find basis using SVD

209 Views Asked by At

Can someone explain to me how should I solve this question?

Suppose

$S = << x_1,...,x_s>> $ $\subseteq$ $R^n$

$T = << y_1,...,y_t>> $ $\subseteq$ $R^n$

Assume that {$x_i$ | i = 1:s} and {$y_i$ | i = 1:t} are linearly independent sets. Use SVD to find a basis for S $\cap $ T.

I'm able to find the orthogonal basis for S and T respectively but I have no idea when it comes to the intersection of both S and T. Thanks in advance.

1

There are 1 best solutions below

0
On

Please provide your partial answer on the orthonormal basis of $S$ and $T$, as it helps to provide an answer in terms you understand.

One of the properties you can use to determine the base for intersection $S \cap T$ is that any base vector of $T$ should preserve norm under multiplication by the normal base of $S$.

So I think you can multiply $V_S$ and $V_T$, limited to the non-zero singular values.