Using the k-nearest neighbor graph for proximity searching in metric spaces

被引:0
|
作者
Paredes, Rodrigo [1 ]
Chavez, Edgar [1 ]
机构
[1] Univ Chile, Ctr Web Res, Dept Comp Sci, Santiago, Chile
关键词
D O I
暂无
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Proximity searching consists in retrieving from a database; objects that are close to a query. For this type of searching problem, the most general model is the metric space, where proximity is defined in terms of a distance function. A solution for this problem consists in building an offline index to quickly satisfy online queries. The ultimate goal is to use as few distance computations as possible to satisfy queries, since the distance is considered expensive to compute. Proximity searching is central to several applications, ranging from multimedia indexing and querying to data compression and clustering. In this paper we present a new approach to solve the proximity searching problem. Our solution is based on indexing the database with the k-nearest neighbor graph (kNNG), which is a directed graph connecting each element to its k closest neighbors. We present two search algorithms for both range and nearest neighbor queries which use navigational and metrical features of the kNNG graph. We show that our approach is competitive against current ones. For instance, in the document metric space our nearest neighbor search algorithms perform 30% more distance evaluations than AESA using only a 0.25% of its space requirement. In the same space, the pivot-based technique is completely useless.
引用
收藏
页码:127 / 138
页数:12
相关论文
共 50 条
  • [1] k-Nearest neighbor searching in hybrid spaces
    Kolbe, Dashiell
    Zhu, Qiang
    Pramanik, Sakti
    INFORMATION SYSTEMS, 2014, 43 : 55 - 64
  • [2] Distributed k-Nearest Neighbor Queries in Metric Spaces
    Ding, Xin
    Zhang, Yuanliang
    Chen, Lu
    Gao, Yunjun
    Zheng, Baihua
    WEB AND BIG DATA (APWEB-WAIM 2018), PT I, 2018, 10987 : 236 - 252
  • [3] Practical construction of k-nearest neighbor graphs in metric spaces
    Paredes, Rodrigo
    Chavez, Edgar
    Figueroa, Karina
    Navarro, Gonzalo
    EXPERIMENTAL ALGORITHMS, PROCEEDINGS, 2006, 4007 : 85 - 97
  • [4] SUBLINEAR TIME APPROXIMATION OF THE COST OF A METRIC k-NEAREST NEIGHBOR GRAPH
    Czumaj, Artur
    Sohler, Christian
    SIAM JOURNAL ON COMPUTING, 2024, 53 (02) : 524 - 571
  • [5] Sublinear time approximation of the cost of a metric k-nearest neighbor graph
    Czumaj, Artur
    Sohler, Christian
    PROCEEDINGS OF THE THIRTY-FIRST ANNUAL ACM-SIAM SYMPOSIUM ON DISCRETE ALGORITHMS (SODA'20), 2020, : 2973 - 2992
  • [6] Efficient k-Nearest Neighbor Searching in Nonordered Discrete Data Spaces
    Kolbe, Dashiell
    Zhu, Qiang
    Pramanik, Sakti
    ACM TRANSACTIONS ON INFORMATION SYSTEMS, 2010, 28 (02)
  • [7] Sublinear time approximation of the cost of a metric k-nearest neighbor graph
    Czumaj, Artur
    Sohler, Christian
    PROCEEDINGS OF THE 2020 ACM-SIAM SYMPOSIUM ON DISCRETE ALGORITHMS, SODA, 2020, : 2973 - 2992
  • [8] Pruned Bi-directed K-nearest Neighbor Graph for Proximity Search
    Iwasaki, Masajiro
    SIMILARITY SEARCH AND APPLICATIONS, SISAP 2016, 2016, 9939 : 20 - 33
  • [9] On k-nearest neighbor searching in non-ordered discrete data spaces
    Kolbe, Dashiell
    Zhu, Qiang
    Pramanik, Sakti
    2007 IEEE 23RD INTERNATIONAL CONFERENCE ON DATA ENGINEERING, VOLS 1-3, 2007, : 401 - +
  • [10] Fast agglomerative clustering using a k-nearest neighbor graph
    Franti, Pasi
    Virmajoki, Olli
    Hautamaki, Ville
    IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2006, 28 (11) : 1875 - 1881