Simple termination conditions for k-nearest neighbor method

Authors:
Mineichi Kudo;Naoto Masuyama;Jun Toyama;Masaru Shimbo
Affiliations:
Division of Systems and Information Engineering, Graduate School of Engineering, Hokkaido University, Kita-13, Nishi-8, Kita-ku, Sapporo 060-8628, Japan;Division of Systems and Information Engineering, Graduate School of Engineering, Hokkaido University, Kita-13, Nishi-8, Kita-ku, Sapporo 060-8628, Japan;Division of Systems and Information Engineering, Graduate School of Engineering, Hokkaido University, Kita-13, Nishi-8, Kita-ku, Sapporo 060-8628, Japan;Faculty of Information Media, Hokkaido Information University, Ebetsu 069-8585, Japan
Venue:
Pattern Recognition Letters
Year:
2003

Citing 12
Cited 0

Model-based object recognition by geometric hashing

ECCV 90 Proceedings of the first european conference on Computer vision
A hashing-oriented nearest neighbor searching scheme

Pattern Recognition Letters
A Fast Algorithm for the Nearest-Neighbor Classifier

IEEE Transactions on Pattern Analysis and Machine Intelligence
Two algorithms for nearest-neighbor search in high dimensions

STOC '97 Proceedings of the twenty-ninth annual ACM symposium on Theory of computing
A Simple Algorithm for Nearest Neighbor Search in High Dimensions

IEEE Transactions on Pattern Analysis and Machine Intelligence
An Algorithm for Finding Best Matches in Logarithmic Expected Time

ACM Transactions on Mathematical Software (TOMS)
Optimal Expected-Time Algorithms for Closest Point Problems

ACM Transactions on Mathematical Software (TOMS)
Multidimensional binary search trees used for associative searching

Communications of the ACM
R-trees: a dynamic index structure for spatial searching

SIGMOD '84 Proceedings of the 1984 ACM SIGMOD international conference on Management of data
Fast Nearest Neighbor Search in High-Dimensional Space

ICDE '98 Proceedings of the Fourteenth International Conference on Data Engineering
The R+-Tree: A Dynamic Index for Multi-Dimensional Objects

VLDB '87 Proceedings of the 13th International Conference on Very Large Data Bases
The X-tree: An Index Structure for High-Dimensional Data

VLDB '96 Proceedings of the 22th International Conference on Very Large Data Bases

Quantified Score

Hi-index	0.10

Visualization

Abstract

The main problem with k-nearest neighbor (k-NN) method is that the computational cost in the search process is proportional to the size of the training samples. Many search algorithms have been proposed to cope with this problem. In this study, we consider some conditions for terminating the search procedure when the true k-NNs have been found in the middle of the search, and we present, as an example, a procedure in the branch-and-bound algorithm. These conditions do not always work for a certain sample, but they reduce the computational cost on average.