
www.Usenet.com
| <-- __Chronological__ --> | <-- __Thread__ --> |
Thanks for the replies. The problem I posted originally is actually much more general than the specific problem I wish to solve. I should have given my specific problem instead of this more general problem: I have a map that takes a full-rank matrix of order m by n (with m<n) to its kernel. The kernel is a linear subspace of dimension n-m in R^n. If we identify each kernel as a point in the range, then the range is a grassman manifold, which I will denote by G(n,n-m). The grassman manifold G(n,n-m) is compact and differentiable, and has dimension m*(n-m). It is equipped with a rotation invariant Riemannian metric, which induces a distance function. Hence we can compute Hausdorff measure. Let S be a subset of G(n,n-m) with zero m*(n-m)-dimensional Hausdorff measure. (To be more precise, it has sigma-finite m*(n-m)-1 dimensional Hausdorff measure.) I know that the preimage of each point in S has Lebesgue measure zero. (In fact the preimage of each point in S has dimension m^2, which is less than the dimension m*n of the domain.) My question is: does the preimage of S have zero Lebesgue measure? Thanks. Simon
| <-- __Chronological__ --> | <-- __Thread__ --> |