东北大学学报:自然科学版 ›› 2015, Vol. 36 ›› Issue (10): 1412-1416.DOI: 10.3969/j.issn.1005-3026.2015.10.010

• 信息与控制 • 上一篇    下一篇

面向组近邻的Top-k空间偏好查询

陈默1, 杨丹2, 谷峪3, 于戈1,3   

  1. (1. 东北大学 计算中心, 辽宁 沈阳110819; 2. 辽宁科技大学 软件学院, 辽宁 鞍山114051; 3. 东北大学 信息科学与工程学院, 辽宁 沈阳110819)
  • 收稿日期:2014-09-30 修回日期:2014-09-30 出版日期:2015-10-15 发布日期:2015-09-29
  • 通讯作者: 陈默
  • 作者简介:陈默(1983-),女,辽宁沈阳人,东北大学讲师,博士; 于戈(1962-),男,辽宁大连人,东北大学教授,博士生导师.`
  • 基金资助:
    国家自然科学基金资助项目(61402093,61402213); 中央高校基本科研业务费专项资金资助项目(N141604001, N120316001).

Top-k Spatial Preference Query for Group Nearest Neighbor

CHEN Mo1, YANG Dan2, GU Yu3, YU Ge1,3   

  1. 1. Computing Center, Northeastern University, Shenyang 110819, China; 2. Software College, University of Science and Technology Liaoning, Anshan 114051, China; 3. School of Information Science & Engineering, Northeastern University, Shenyang 110819, China.
  • Received:2014-09-30 Revised:2014-09-30 Online:2015-10-15 Published:2015-09-29
  • Contact: YANG Dan
  • About author:-
  • Supported by:
    -

摘要: 空间偏好查询是当前空间查询研究中的一类热点问题,而现有的空间偏好查询不能有效支持面向组用户的位置服务应用.为此,提出一类新型空间偏好查询——面向组近邻的Top-k空间偏好查询 (Top-k spatial preference query for group nearest neighbor).该查询通过查找特征对象的λ子集组近邻最终为用户返回评分值最高的前k个λ子集.为了高效执行这一查询,给出了两种查询算法:TSPQ-G及TSPQ-G*.其中TSPQ-G*在TSPQ-G的基础上,通过空间剪枝及高效的特征对象索引树遍历策略大幅减少I/O代价,进而有效提高了该查询的执行效率.实验采用多个数据集验证了所提算法在不同参数设置下的有效性.

关键词: 空间偏好, 位置服务, 组近邻, 剪枝, 查询

Abstract: Spatial preference query is a popular focus of the current research on spatial queries. However, the present spatial preference queries cannot be used in the location-based services for group users. To solve this problem, a novel type of spatial preference query, namely, Top-k spatial preference query for group nearest neighbor (TSPG) was proposed, which retrieves the k λ-subsets with the highest score through finding λ-subsets group nearest neighbors of the feature objects. Two algorithms, namely, TSPQ-G and TSPQ-G* were designed for efficient query processing. Based on the TSPQ-G, the TSPQ-G* was developed by performing spatial pruning strategies and efficient traversal strategies of feature objects index, which effectively reduces I/O cost and improves query efficiency. Experimental results on several datasets demonstrated the effectiveness of the proposed algorithms for different setups.

Key words: spatial preference, location-based service, group nearest neighbor, pruning, query

中图分类号: