[Math] Nearest neighbors search

[Gilles <gi...@harfang.homelinux.org>]

Hi.

Referring to the newly created MATH-1235.

I was looking for a way to improve the performance of a code that also
needs, for some point "p", to select the "n" closest neighbours (from
a list of sample data points).

Browsing through web pages, some refer to BSP as a generalization of
kd-tree; hence the question whether the implementation in CM can be
used to provide the functionality in a way that is independent of its
projected use to resolve MATH-1235.

What do you think?


Gilles


---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org

Re: [Math] Nearest neighbors search

[Thomas Neidhart <th...@gmail.com>]

On Fri, Jun 19, 2015 at 3:56 PM, Gilles <gi...@harfang.homelinux.org>
wrote:

> Hi.
>
> Referring to the newly created MATH-1235.
>
> I was looking for a way to improve the performance of a code that also
> needs, for some point "p", to select the "n" closest neighbours (from
> a list of sample data points).
>
> Browsing through web pages, some refer to BSP as a generalization of
> kd-tree; hence the question whether the implementation in CM can be
> used to provide the functionality in a way that is independent of its
> projected use to resolve MATH-1235.
>
> What do you think?
>

I do not know if the BSPTree can be adapter for this purpose.

In the case of DBSCAN I need an n-dimensional KD tree, so the BSPTree
solution would not work for me at least.

btw. scipy also offers a Kdtree as part of its spatial package.

Thomas