In Nearest Neighbours are trees other than BruteTree worth it for sets with between n=2^3 and n=2^20ish for KNN in 2D cartesian space

kristoffer.carlsson · March 28, 2022, 3:58pm

Yes, KDTrees are good in low dimensions and get (significantly) worse with higher dimensions. Using a proper tree should be a big win. But just benchmarking should show the truth for your specific use case.

Topic		Replies	Views
How to compute neighbour network using NearestNeighbors.jl Specific Domains	4	1538	January 23, 2017
KDTree in NearestNeighbors.jl with Rectangles as leaf nodes Specific Domains	1	459	March 26, 2019
[ANN] VPTrees.jl - nearest neighbor search in metric spaces Package Announcements	20	2751	January 30, 2020
Suggestion needed: Tree like structure for adaptive binning of multidim data Statistics	6	598	October 2, 2020
K nearest neighbor utility in Julia Community	13	3961	February 3, 2021

In Nearest Neighbours are trees other than BruteTree worth it for sets with between n=2^3 and n=2^20ish for KNN in 2D cartesian space

Related topics