Please use this identifier to cite or link to this item:
|Title:||SiMPSON: Efficient similarity search in metric spaces over P2P structured overlay networks||Authors:||Vu, Q.H.
|Issue Date:||2009||Citation:||Vu, Q.H.,Lupu, M.,Wu, S. (2009). SiMPSON: Efficient similarity search in metric spaces over P2P structured overlay networks. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) 5704 LNCS : 498-510. ScholarBank@NUS Repository. https://doi.org/10.1007/978-3-642-03869-3_48||Abstract:||Similarity search in metric spaces over centralized systems has been significantly studied in the database research community. However, not so much work has been done in the context of P2P networks. This paper introduces SiMPSON: a P2P system supporting similarity search in metric spaces. The aim is to answer queries faster and using less resources than existing systems. For this, each peer first clusters its own data using any off-the-shelf clustering algorithms. Then, the resulting clusters are mapped to one-dimensional values. Finally, these one-dimensional values are indexed into a structured P2P overlay. Our method slightly increases the indexing overhead, but allows us to greatly reduce the number of peers and messages involved in query processing: we trade a small amount of overhead in the data publishing process for a substantial reduction of costs in the querying phase. Based on this architecture, we propose algorithms for processing range and kNN queries. Extensive experimental results validate the claims of efficiency and effectiveness of SiMPSON. © 2009 Springer.||Source Title:||Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)||URI:||http://scholarbank.nus.edu.sg/handle/10635/41146||ISBN:||3642038689||ISSN:||03029743||DOI:||10.1007/978-3-642-03869-3_48|
|Appears in Collections:||Staff Publications|
Show full item record
Files in This Item:
There are no files associated with this item.
checked on Sep 20, 2022
checked on Sep 22, 2022
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.