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

[lozhnikov]

I think the second option will lead to enormous duplication of the code. Right now, `VantagePointTree` has almost the same code as `BinarySpaceTree`. And if I am not mistaken we have to add the third similar class in order to implement that.

I guess it is better to add a method in each `TreeType`: `IsFirstPointCentroid()`. That should not lead to any slowdown since this check is used rarely. (There are two options: we can implement this method in the class definition, so the method will be inlined or we can add two tree traits `AllFirstPointsAreCentroids` and `SomeFirstPointsAreCentroids` and invoke this method when it is needed. I think the first option is better.)

Maybe the fourth option is better. We can try to implement the third approach and test performance in order to choose between the third and the fourth options.

So, I guess we should choose either the third or the fourth option.

---
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-232144887
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.cc.gatech.edu/pipermail/mlpack-git/attachments/20160712/19b96e3c/attachment.html>