Revision history [back]
 | 1 | initial version |
Looking at the code (with ??), J.eigenspaces() is equivalent to J.adjacency_matrix(vertices=list(J)).right_eigenspaces(), which uses another ordering of the vertices. Since the adjacency matrices are different (but similar), they produce the same eigenvalues, but different eigenspaces. You can use
sage: J.adjacency_matrix().is_similar(J.adjacency_matrix(vertices=list(J)), transformation=True)[1]
To get the change of basis between the two.
| | 2 | No.2 Revision |
Looking at the code (with ??), J.eigenspaces() is equivalent to J.adjacency_matrix(vertices=list(J)).right_eigenspaces(), which uses another ordering of the vertices. Since the adjacency matrices are different (but similar), they produce the same eigenvalues, but different eigenspaces. You can use
sage: J.adjacency_matrix().is_similar(J.adjacency_matrix(vertices=list(J)), transformation=True)[1]
To get the change of basis between the two.
I've never worked with graph eigenspace before, is it supposed to be independent of the ordering ?
