Jeff Bradberry
Jeff Bradberry
> I might be getting lost in the sauce a little here, but some more comments. > > * Perhaps most importantly: isn't the tree centroid the same thing as...
Hm. ``` In [42]: for g in nx.graph_atlas_g(): ...: if len(g.nodes) == 0: ...: continue ...: if not nx.is_tree(g): ...: continue ...: b = nx.barycenter(g) ...: c = nx.tree_centroid(g) ...:...
A bit of ambiguous argument: I've been trying systematically generated small trees (I think I got up to 9 nodes before it was taking too long for my taste) and...
What the heck is going on with test/dispatch (3.11)? It's throwing a bunch of `KeyError: 'G'` for my tests.
S ( u ) = S ( c ) − s + ( N − s ) = S ( c ) + N − 2 s : I'm pretty...
My attempted proof: Consider a node $a$ in tree $T$. What are the conditions where moving along the edge $ab$ will improve the barycenter objective function, but increase the weight?...
> I might be missing something, but aren't we saying the same thing? I ~might~ almost certainly have misread what you meant. I should not be attempting to read or...
Having said all of this I don't think I oppose the proposed use, but hope that this can be done in a follow-on PR instead of added to this one.
> I'm not sure I understand this part of your proof: > > > In the former case, it is sufficient for > > > > N_A > N_B >...
All right, I think I've got them all.