A novel high-dimensional index method based on the mathematical features

Yu Zhang, Jiayu Li*, Ye Yuan

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Nowadays the nearest neighbor (NN) search in the high dimensional space can be applied in many fields and it becomes the focus of information science. Usually, R-near neighbor that sets a fixed query range R is used in place of NN search. However, the traditional methods for R-near neighbor can not achieve the satisfactory performance in the high dimensional space due to the curse of dimensionality. Moreover, some methods is based on probabilistic guarantees so it does not provide the 100% accuracy guarantee. To improve the problem, in this paper, we propose a novel idea to build the index structure. This method is based on the mathematical features of the coordinates of the data points. Specifically, we employ the mean value and the standard deviation of the coordinate to index the data point. This method can efficiently solve the R-NN search with the 100% accuracy guarantee in the high dimensional space. Extensive experimental results demonstrate the effectiveness of the proposed methods.

Original languageEnglish
Title of host publicationBig Data Computing and Communications - 2nd International Conference, BigCom 2016, Proceedings
EditorsYu Wang, Ge Yu, Guoren Wang, Yanyong Zhang, Zhu Han
PublisherSpringer Verlag
Pages257-271
Number of pages15
ISBN (Print)9783319425528
DOIs
Publication statusPublished - 2016
Externally publishedYes
Event2nd International Conference on Big Data Computing and Communications, BigCom 2016 - Shenyang, China
Duration: 29 Jul 201631 Jul 2016

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume9784
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference2nd International Conference on Big Data Computing and Communications, BigCom 2016
Country/TerritoryChina
CityShenyang
Period29/07/1631/07/16

Keywords

  • High-dimension
  • Multimedia
  • R-near neighbor

Fingerprint

Dive into the research topics of 'A novel high-dimensional index method based on the mathematical features'. Together they form a unique fingerprint.

Cite this