[mlpack-git] [mlpack/mlpack] Universal B tree implementation (#746)

[Ryan Curtin]

I need to review the paper before I can make more comments, but I am surprised that this is so much slower than the kd-tree.  I am seeing performance about the same as the rp-tree.  To me this seems odd.  I can understand why the vp-tree is slow and I can understand why the rp-tree (and variants) are slow, and I know that the R tree and its variants will also be slow as compared to the kd-tree, but I had thought that the ub-tree, since it was so similar with axis-aligned pretty-tight hyperrectangles, would be much faster.  Even with low-dimensional data, it often takes an order of magnitude more base cases.

When I am able to review the paper, maybe I will be able to reason if something is wrong.  A cursory inspection of the code to calculate the minimum and maximum distances seems fine, but the only hesitation I have is about whether or not the actual strategy of splitting points into bounds is correct.  (But again I have to refer to the paper before I can comment more there.)

Outside of that, the code quality seems great.

-- 
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/mlpack/mlpack/pull/746#issuecomment-240315737
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.cc.gatech.edu/pipermail/mlpack-git/attachments/20160816/6b1d7d31/attachment.html>