Dimensional Testing for Reverse k-Nearest Neighbor Search

被引:25
作者
Casanova, Guillaume [1 ]
Englmeier, Elias [2 ]
Houle, Michael E. [3 ]
Kroeger, Peer [2 ]
Nett, Michael [4 ]
Schubert, Erich [5 ]
Zimek, Arthur [6 ]
机构
[1] Off Natl Etud & Rech Aerosp, DCSD, Palaiseau, France
[2] Ludwig Maximilians Univ Munchen, Munich, Germany
[3] NII, Tokyo, Japan
[4] Google Japan, Tokyo, Japan
[5] Heidelberg Univ, Heidelberg, Germany
[6] Univ So Denmark, Odense, Denmark
来源
PROCEEDINGS OF THE VLDB ENDOWMENT | 2017年 / 10卷 / 07期
关键词
QUERIES;
D O I
10.14778/3067421.3067426
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Given a query object q, reverse k-nearest neighbor (RkNN) search aims to locate those objects of the database that have q among their k-nearest neighbors. In this paper, we propose an approximation method for solving RkNN queries, where the pruning operations and termination tests are guided by a characterization of the intrinsic dimensionality of the data. The method can accommodate any index structure supporting incremental (forward) nearest-neighbor search for the generation and verification of candidates, while avoiding impractically-high preprocessing costs. We also provide experimental evidence that our method significantly outperforms its competitors in terms of the tradeoff between execution time and the quality of the approximation. Our approach thus addresses many of the scalability issues surrounding the use of previous methods in data mining.
引用
收藏
页码:769 / 780
页数:12
相关论文
共 51 条
[1]   Online hierarchical clustering in a data warehouse environment [J].
Achtert, E ;
Böhm, C ;
Kriegel, HP ;
Kröger, P .
FIFTH IEEE INTERNATIONAL CONFERENCE ON DATA MINING, PROCEEDINGS, 2005, :10-17
[2]  
Achtert E., 2006, P SIGMOD, P515, DOI DOI 10.1145/1142473.1142531
[3]  
Achtert E, 2006, CIKM, P788
[4]   Estimating Local Intrinsic Dimensionality [J].
Amsaleg, Laurent ;
Chelly, Oussama ;
Furon, Teddy ;
Girard, Stephane ;
Houle, Michael E. ;
Kawarabayashi, Ken-ichi ;
Nett, Michael .
KDD'15: PROCEEDINGS OF THE 21ST ACM SIGKDD INTERNATIONAL CONFERENCE ON KNOWLEDGE DISCOVERY AND DATA MINING, 2015, :29-38
[5]  
[Anonymous], BTW
[6]  
[Anonymous], 2005, P 31 INT C VER LARG
[7]  
[Anonymous], MMCBIR
[8]  
[Anonymous], 2000, ACM SIGMOD Workshop on Research Issues in Data Mining and Knowledge Discovery
[9]  
Bache K., 2013, UCI Machine Learning Repository
[10]  
Beygelzimer A, 2006, P 23 INT C MACH LEAR, P97, DOI DOI 10.1145/1143844.1143857