【作者】 刘健；

【导师】 马宗民；

【作者基本信息】 东北大学 ， 计算机应用技术， 2014， 博士

【摘要】 随着Web技术的广泛应用,互联网已成为人们获取信息的主要工具。由于互联网数据格式通常存在着异构性,为了高效地处理Web数据信息,越来越多的应用系统采用了XML (Extensible Markup Language)作为数据表示和交换的标准。作为下一代的Web语言,XML具有格式独立、与平台无关的特性。XML技术在当前的互联网络环境中扮演着越来越重要的角色,它事实上已经成为了Web时代实现数据信息智能管理的基石。正是在这样的应用背景下,基于XML的数据管理应用引起了学术界和工业界的高度关注,有关XML数据存储与查询的理论和技术也相应地成为了数据库领域中的研究热点。现实世界应用中存在着大量的不精确和不确定信息,在数据库中融入模糊数据处理技术以满足现实需求成为了当前研究热点,并因此产生了有关模糊XML数据管理的需求。然而,从现有的有关模糊XML的研究文献来看,该领域研究成果还很少,特别对模糊XML数据的存储以及查询等问题的研究还处于起步阶段,许多重要的理论问题和技术难点当前仍是研究空白。鉴于这样的研究现状,本论文的研究工作将从构建模糊XML数据库系统的角度出发,围绕模糊XML数据存储与查询问题深入展开,本文将提出一套行之有效的模糊XML数据管理的解决方案。具体研究内容包括以下几个方面：(1)针对模糊XML数据的存储问题,提出了基于关系数据库的存储方法,同时提出了从模糊XML模型到UML概念数据模型的映射转换方法。首先研究了模式信息未知情况下模糊XML数据的关系数据库存储问题,提出了基于XML边信息的关系数据库存储方法,给出了从XML查询表达式到SQL查询表达式的映射转换方法。通过边位置序号的引入,有效解决了现有方法在数据转换时出现的语义次序信息丢失问题。同时研究了从模糊XML模型到UML数据模型的转换问题,提出了模式信息未知情况下模糊XML模型到UML概念数据模型的映射以及模式构建方法,解决了现有方法在进行模式转换时需要强制系统提供XML模式信息的问题。(2)针对模糊XML数据的查询问题,提出了基于小枝模式的模糊XML数据查询方法。基于整体匹配策略,研究了模糊XML小枝模式的结构化查询问题,提出了复杂逻辑语义下,特别是包含逻辑或和逻辑非语义的小枝匹配问题的解决方案,解决了现有方法在处理逻辑或和逻辑非语义的小枝模式查询时需要重复读取源数据所带来的读取代价过大问题,提升了小枝模式匹配的处理效率。基于孩子线索,定义了异构哈希表,并在此基础上研究了异构XML数据源下的小枝匹配方法,解决了现有方法需要预先集成异构文档再计算小枝匹配所带来的集成代价过大问题,提升了异构模糊XML小枝模式的查询性能。(3)针对严格结构化查询可能返回过少甚至空查询结果的问题,提出了基于语义相似度的针对模糊XML数据的近似查询方法。首先,基于用户查询和数据分布推测用户查询意图,据此提出了模糊XML数据查询的权重评估方法。然后,分别给出了查询结构相似度以及查询内容语义相似度评估方法。在此基础上,提出了基于模糊XML数据的自适应查询松弛方法以及结果排序方法,解决了现有结构化查询方法在响应查询时需要强制用户了解数据模式信息并提供严格查询语言语法所带来的用户查询负担过重问题,提高了模糊XML近似查询的查准率和查全率。更多还原

【Abstract】 With the wide application of Web technologies, Internet has become an important tool for information communication. Due to its heterogeneity of data on the Internet, more and more application systems choose XML (Extensible Markup Language) as the de-facto standard for information representation and exchange. As the next generation of the Web language, XML offers particular advantages such as extensible, platform-independent, etc. As long as applications support XML, a seamless exchange of information among these applications would be provided. XML plays a signifincat role in a Web environment, and it has become the cornerstone of intelligent management systems. Under such a background, the problem of effective XML data management has received significant attention from both academic and industrial communities, and theories and techniques of XML data storages and queries have become hot topics of database researchers.Information imprecision and uncertainty exist in many real-world applications. For this reason, a lot of interests in the research of fuzzy data managements have been stimulated. A combination of databases with fuzzy data management techniques has become the focus of research, which also creates a new set of uncertain data management requirements involving XML. Unfortunately, although some researches have devoted to investigate uncertain XML data managements in recent years, the research of fuzzy XML is still in its infant stage. Relative little work has been carried out in storing and querying XML data towards the representation of imprecise and uncertain concepts. In view of this situation, this paper proposes an effective and efficient solution and related techniques along with fuzzy data storages and queries for constructing a robust system for XML data managements. The main contributions of this paper are as follows:Firstly, to deal with the problem of fuzzy XML data storages, the issues of stroing fuzzy XML data in relational databases, and reengineering fuzzy XML in the UML data model, are investigated. In particular, an edge-based mapping approach and a query transformation approach are proposed to shred fuzzy XML data into relational data and transform XML query expressions to SQL expressions respectively. The proposed approaches effectively avoid losing semantic order information during the reengineering from the fuzzy XML data model to the relational data model. Moreover, a rule-based mapping approach is developed to reengineer fuzzy XML in the UML data model. The proposed approach avoids requiring systems to provide XML schemas during the mapping from the fuzzy XML data model to the UML data model.Secondly, based on the fuzzy labeling scheme, the issue of twig pattern queries with complex predicates is investigated. By using the holistic matching strategy, efficient approaches for answering twig pattern queries with complex predicates such as OR or NOT connectives in homogeneous fuzzy XML documents are proposed. The proposed methods avoid the re-scan problem and promote the query performance, when producing the matches. Holistic approach for twig matching in heterogeneous fuzzy XML documents is also discussed in the thesis. The proposed method avoids requiring a significant cost to integrate heterogeneous documents beforehand, and it promotes the query performance when producing the matches in heterogeneous fuzzy XML documents.Finally, to deal with the problem of few or empty answers returned by using structured queries (twig pattern queries), in response to a user query, an adaptive approximate query approach, which is based on semantic similarities, is proposed. By analyzing data distribution and speculating the factors that users are more concerned about, a weight assignment method is proposed. Then, evaluation methods for structure and content similaries are proposed, respectively. On this basis, an adaptive query relaxation approach and a ranking method are presented. The proposed approach avoids requiring users to understand the XML schemas and provide rigid query expressions, which effectively reduces the query cost, and improves the precisions and recalls.更多还原