[mlpack-git] [mlpack/mlpack] Vantage point tree (#708)

[Ryan Curtin]

This might be a clue to the performance issue.  I modified the code to print information on the bounds and distances for single-tree search.  Here's what I see on the first node of the tree, for query point 0 (with the corel dataset, but you could reproduce these results with probably any dataset):

```
[DEBUG] MinDistance() with inner radius 0, outer radius 8.48376 and center distance 0 (innerDistance 0, outerDistance 0).
[INFO ] Score query index 0 and reference node 1c20746 (child 1): 0 vs. bound 1.79769e+308.
[DEBUG] MinDistance() with inner radius 0, outer radius 13.8071 and center distance 0 (innerDistance 0, outerDistance 0).
[INFO ] Score query index 0 and reference node 20747c17002 (child 2): 0 vs. bound 1.79769e+308.
```

It's expected that neither the first inner child nor the first outer child should be pruned.  But what's not expected is that my code in `HollowBallBound` (the `[DEBUG]` message) is reporting that the inner radius of the outer child is 0!  Maybe there is some bug where that value is calculated?

---
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/708#issuecomment-235101060
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.cc.gatech.edu/pipermail/mlpack-git/attachments/20160725/da7990f8/attachment.html>