0

I was wondering, because I can't find anything about a spacial partionning tree that would be bottom up built:

Is it possible (I guess yes but with what efficiency?)
Does it exist in practical case?
Any documentation about this?

Is this question senseless?

EDIT:
by spacial partitioning, I mean tree similar to BSP, Quad Tree, kd-tree, AABB-tree... But it seems to me that theses trees are build from large area to smaller ones (so top-down).

My question was:
Are there bottom up algorithm to achieve similar algorithm?

shapiro yaacov
  • 2,308
  • 2
  • 26
  • 39
Krapow
  • 611
  • 6
  • 26
  • Please explain what you mean by *Spacial partitioning*. In any case, I'm going to guess maybe this can help: https://en.wikipedia.org/wiki/R-tree – shapiro yaacov Feb 04 '16 at 14:15
  • @shapiro yaacov : Edited the question to be more precise – Krapow Feb 04 '16 at 15:17
  • Why would you care how it's build? What prevents you from joining smaller ones into bigger ones? – Sorin Feb 04 '16 at 15:23
  • @Sorin : Why would I care? Because I am just curious. What prevents me from joining smaller ones into bigger ones? Because I don't know how to do that, I can't figure out an algorithm, that was the question. – Krapow Feb 04 '16 at 15:25
  • 1
    Please read this part: https://en.wikipedia.org/wiki/R-tree#Bulk-loading. In bulk loading it seems that it IS build bottom up... – shapiro yaacov Feb 04 '16 at 15:36
  • curios - nothing wrong with that. For a QuadTree bottom up would mean that you construct 4 adjacent cells then unite them into a node above (and similar would work for most other structures). For this to be efficient you must get the elements in the right order, which is kinda hard to describe and very structure specific. Since the usual algorithm are O(NLogN) anyway it's probably not worth the trouble. If you have more details about the problem maybe somebody would point to more specific data structure or algorithm. – Sorin Feb 04 '16 at 16:38

0 Answers0