孟宪福,孟泓汐,张振强.基于区位码和距离的Chord网络高维数据范围检索[J].,2011,(3):433-439 |
基于区位码和距离的Chord网络高维数据范围检索 |
Bit-code and distance-based range query for high-dimensional data in Chord network |
|
DOI:10.7511/dllgxb201103021 |
中文关键词: P2P 范围检索 iDistance 位置保持哈希函数 区位码 |
英文关键词: P2P range query iDistance locality-preserving hashing function bit-code |
基金项目: |
|
摘要点击次数: 1228 |
全文下载次数: 1075 |
中文摘要: |
结构化对等网络对数据的范围检索等相似查询缺乏有效的支持.由于利用iDistance索引进行范围查询时会引入很多“误中点”,提出了一种利用区位码和距离来建立高维数据索引的方法.该方法首先利用Code-Distance技术来建立高维数据的一维索引,并采用位置保持哈希函数给予每个索引值一个唯一标识,该标识将被保存在Chord环节点上,从而生成BM-Chord系统.同时,详细叙述了数据过滤技术和范围查询策略.模拟实验结果表明,BM-Chord系统在减小中间结果集大小和提高查全率等方面是有效的. |
英文摘要: |
A structured P2P network is not suitable for similarity search, such as range query. Since a lot of false hits may be obtained by using iDistance indexing technique, a high-dimensional data indexing algorithm based on bit-code and distance is proposed. The Code-Distance technique and locality-preserving hashing function are utilized to respectively construct one-dimensional index from data with high-dimension and assign a unique identifier for each index, and those identifiers are kept into nodes to establish BM-Chord network. The data filtering technique and range query strategy are described in detail as well. The simulation results testify the effectiveness of BM-Chord network in terms of the intermediate data set and recall ratio. |
查看全文
查看/发表评论 下载PDF阅读器 |
关闭 |
|
|
|