arXivDaily arXiv每日学术速递 周一至周五更新
arXiv cs/0310035cs.DB
收藏

在以数据库为中心的Web应用中支持探索性搜索

Supporting Exploratory Queries in Database Centric Web Applications

Abhijit Kadlag, Amol Wanjari, Juliana Freire, Jayant R. Haritsa

摘要:

以数据库为中心的Web应用程序的用户,特别是在电子商务领域,经常求助于探索性的"试错“查询,因为底层数据空间是巨大的和不熟悉的,并且在这个空间中有几个搜索属性的替代方案。例如,寻找便宜的机票通常涉及提出多个查询,改变航班时间,日期和机场位置。从用户和服务器的角度来看,探索性查询都是有问题的。对于数据库服务器,它会导致有效吞吐量的急剧减少,因为在每个连续的查询中重复了许多处理。对于客户端来说,这会导致响应时间显著增加,尤其是在通过无线信道访问服务时。在本文中,我们研究的自动化技术的设计,以尽量减少重复的探索性查询的需要。具体来说,我们提出了SAUNA,服务器端查询松弛算法,给定用户的初始范围查询和所需的基数的答案集,产生一个宽松的查询,预计将包含所需数量的答案。该算法采用了范围查询特定的距离度量,加权产生所需形状的宽松的查询(例如,纵横比保持),并利用多维直方图的查询大小估计。一个详细的性能评估的SAUNA在各种多维数据集表明,其宽松的查询可以显着降低成本与探索性查询处理。

英文摘要:

Users of database-centric Web applications, especially in the e-commerce domain, often resort to exploratory ``trial-and-error'' queries since the underlying data space is huge and unfamiliar, and there are several alternatives for search attributes in this space. For example, scouting for cheap airfares typically involves posing multiple queries, varying flight times, dates, and airport locations. Exploratory queries are problematic from the perspective of both the user and the server. For the database server, it results in a drastic reduction in effective throughput since much of the processing is duplicated in each successive query. For the client, it results in a marked increase in response times, especially when accessing the service through wireless channels. In this paper, we investigate the design of automated techniques to minimize the need for repetitive exploratory queries. Specifically, we present SAUNA, a server-side query relaxation algorithm that, given the user's initial range query and a desired cardinality for the answer set, produces a relaxed query that is expected to contain the required number of answers. The algorithm incorporates a range-query-specific distance metric that is weighted to produce relaxed queries of a desired shape (e.g. aspect ratio preserving), and utilizes multi-dimensional histograms for query size estimation. A detailed performance evaluation of SAUNA over a variety of multi-dimensional data sets indicates that its relaxed queries can significantly reduce the costs associated with exploratory query processing.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑