Temporally relevant parallel top-k spatial keyword search

IF 1.8 Q2 GEOGRAPHY
S. Ray, B. Nickerson
{"title":"Temporally relevant parallel top-k spatial keyword search","authors":"S. Ray, B. Nickerson","doi":"10.5311/josis.2022.24.199","DOIUrl":null,"url":null,"abstract":"New spatio-textual indexing methods are needed to support efficient search and update of the massive amounts of spatially referenced text being generated. Location based services using geo-tagged documents provide valuable ranked recommendations about nearby restaurants, services, sales, emergency events, and visitor attractions. Consequently, top-k spatial keyword search queries (TkSKQ) have received a lot of attention from the research community. Several spatio-textual indexes have been proposed to efficiently support TkSKQ. Some of these indexes support updates based on live document streams, but the ranking schemes employed by them do not simultaneously incorporate temporal relevance, textual similarity and spatial proximity. Moreover, existing approaches have limited or no capability to exploit parallelism with document ingestion and query execution. We present a parallel spatio-textual index, Pastri, to address the aforementioned issues. Pastri can be updated incrementally over real-time spatio-textual document streams. To support temporally relevant ranking of continuously generated document streams, we propose a dynamic ranking scheme. Our approach retrieves the top-k documents that are most temporally relevant at the time of a query execution. We implemented Pastri and we integrate it within a system with a persistent document store and several thread pools to exploit parallelism at various levels. Experimental evaluation involving real-world datasets and synthetic datasets (that we created) demonstrates that our system is able to sustain high document update throughput. Furthermore, Pastri's TkSKQ search performance is one to two orders of magnitude faster than other spatio-textual indexes.","PeriodicalId":45389,"journal":{"name":"Journal of Spatial Information Science","volume":" ","pages":""},"PeriodicalIF":1.8000,"publicationDate":"2022-06-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of Spatial Information Science","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.5311/josis.2022.24.199","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"GEOGRAPHY","Score":null,"Total":0}
引用次数: 1

Abstract

New spatio-textual indexing methods are needed to support efficient search and update of the massive amounts of spatially referenced text being generated. Location based services using geo-tagged documents provide valuable ranked recommendations about nearby restaurants, services, sales, emergency events, and visitor attractions. Consequently, top-k spatial keyword search queries (TkSKQ) have received a lot of attention from the research community. Several spatio-textual indexes have been proposed to efficiently support TkSKQ. Some of these indexes support updates based on live document streams, but the ranking schemes employed by them do not simultaneously incorporate temporal relevance, textual similarity and spatial proximity. Moreover, existing approaches have limited or no capability to exploit parallelism with document ingestion and query execution. We present a parallel spatio-textual index, Pastri, to address the aforementioned issues. Pastri can be updated incrementally over real-time spatio-textual document streams. To support temporally relevant ranking of continuously generated document streams, we propose a dynamic ranking scheme. Our approach retrieves the top-k documents that are most temporally relevant at the time of a query execution. We implemented Pastri and we integrate it within a system with a persistent document store and several thread pools to exploit parallelism at various levels. Experimental evaluation involving real-world datasets and synthetic datasets (that we created) demonstrates that our system is able to sustain high document update throughput. Furthermore, Pastri's TkSKQ search performance is one to two orders of magnitude faster than other spatio-textual indexes.
时间相关的并行top-k空间关键字搜索
需要新的空间文本索引方法来支持对生成的大量空间引用文本的高效搜索和更新。使用地理标记文档的基于位置的服务提供有关附近餐馆、服务、销售、紧急事件和旅游景点的有价值的排名推荐。因此,top-k空间关键字搜索查询(TkSKQ)受到了研究界的广泛关注。为了有效地支持TkSKQ,提出了几种空间文本索引。其中一些索引支持基于实时文档流的更新,但是它们采用的排序方案没有同时考虑时间相关性、文本相似性和空间接近性。此外,现有的方法在利用文档摄取和查询执行的并行性方面能力有限,甚至没有。我们提出了一个平行的空间-文本索引,Pastri,以解决上述问题。Pastri可以在实时空间文本文档流上增量更新。为了支持对连续生成的文档流进行时间相关排序,我们提出了一种动态排序方案。我们的方法检索在查询执行时最具临时相关性的前k个文档。我们实现了Pastri,并将其集成到一个具有持久文档存储和几个线程池的系统中,以利用不同级别的并行性。涉及真实世界数据集和合成数据集(我们创建的)的实验评估表明,我们的系统能够维持高文档更新吞吐量。此外,Pastri的TkSKQ搜索性能比其他空间文本索引快一到两个数量级。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 求助全文
来源期刊
CiteScore
5.10
自引率
0.00%
发文量
5
审稿时长
9 weeks
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信