Exploiting Sample-Data Distributions to Reduce the Cost of Nearest-Neighbor Searches with Kd-Trees

Authors:
Douglas A. Talbert;Douglas H. Fisher
Affiliations:
-;-
Venue:
IDA '99 Proceedings of the Third International Symposium on Advances in Intelligent Data Analysis
Year:
1999

Citing 8
Cited 0

Voronoi diagrams—a survey of a fundamental geometric data structure

ACM Computing Surveys (CSUR)
A Review and Empirical Evaluation of Feature Weighting Methods for aClass of Lazy Learning Algorithms

Artificial Intelligence Review - Special issue on lazy learning
Voronoi diagrams and Delaunay triangulations

Handbook of discrete and computational geometry
Data structures and algorithms for nearest neighbor search in general metric spaces

SODA '93 Proceedings of the fourth annual ACM-SIAM Symposium on Discrete algorithms
An Algorithm for Finding Best Matches in Logarithmic Expected Time

ACM Transactions on Mathematical Software (TOMS)
Multidimensional divide-and-conquer

Communications of the ACM
OPT-KD: An Algorithm for Optimizing Kd-Trees

ICML '99 Proceedings of the Sixteenth International Conference on Machine Learning
Multiresolution instance-based learning

IJCAI'95 Proceedings of the 14th international joint conference on Artificial intelligence - Volume 2

Quantified Score

Hi-index	0.00

Visualization

Abstract

We present KD-DT, an algorithm that uses a decision-tree-inspired measure to build a kd-tree for low cost nearest-neighbor searches. The algorithm starts with a "standard" kd-tree and uses searches over a training set to evaluate and improve the structure of the kd-tree. In particular, the algorithm builds a tree that better insures that a query and its nearest neighbors will be in the same subtree(s), thus reducing the cost of subsequent search.