[mlpack-git] [mlpack/mlpack] Hilbert R tree (#664)

[Ryan Curtin]

Overall this looks good, but there are a couple points we should figure out:

 * If the recursive strategy is very slow, do you think it is salvageable, or do you think we should forget it and just go with the discrete strategy?  I think there are opportunities for speedups in both DiscreteHilbertValue and RescursiveHilbertValue but I have not figured out the exact details of how that can be done, and I am not sure if even with those speedups, RecursiveHilbertValue would be competitive with DiscreteHilbertValue.

 * Do you have any tests for DiscreteHilbertValue or RecursiveHilbertValue?  The code there is pretty complex, we should test it in a standalone manner to make sure it is right.  I didn't see any tests for either class, though.

The refactoring of the auxiliary information is really nice, `RectangleTree` is already way nicer than when you started working on it. :)

---
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/664#issuecomment-223461713
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.cc.gatech.edu/pipermail/mlpack-git/attachments/20160602/c49eaddf/attachment.html>