Effective approach for an extended P-skyline query

Issue title: ICNC-FSKD 2015

Guest editors: Zheng Xiao and Kenli Li

Article type: Research Article

Authors: Zhou, Xu^a | Zhou, Yantao^{a; b; *} | Xiao, Guoqing^a | Zeng, Yifu^b | Zheng, Fei^b

Affiliations: [a] School of Information Science and Engineering, Hunan University, Changsha, China | [b] College of Electrical and Information Engineering, Hunan University, Changsha, China

Correspondence: [*] Corresponding author. Yantao Zhou. E-mail: yantao_z@hun.edu.cn.

Abstract: With the rapid growth of uncertain data available in many real life applications, a probabilistic skyline query, namely P-skyline query, has been developed and received widespread concern. However, the P-skyline query usually reports results, which have dominant relationship. This contradicts with the incomparable property of skyline queries. Motivated by this, we extend the P-skyline query and formulate an EP-skyline (EPS) query. Thereafter, to develop the processing performance of EPS query, we utilize an index, PR-tree, to organize uncertain datasets and employ efficient pruning strategies to reduce the search space. Moreover, an effective algorithm is developed for the EPS query. Extensive experiments verify that our EPS query could always return better query results than P-skyline query with much less CPU cost, I/O cost and memory cost.

Keywords: Data management, skyline query, uncertain data

DOI: 10.3233/JIFS-169016

Journal: Journal of Intelligent & Fuzzy Systems, vol. 31, no. 2, pp. 849-858, 2016