关联课程数据构建及存储方法研究
发布时间：2019-03-30 23:16
【摘要】：关联数据的概念于2006年被提出。用关联数据技术发布数据信息,是实现数据万维网最重要的一步。W3C组织公布的资源描述框架和Web本体语言对文档中出现的概念和概念间的关系做了形式化的定义,使得网络数据具有了全球通用性和机器可读性的特点。关联数据的最主要作用是关联创建和数据整合。同时,关联数据技术对电子学习系统也提供了富有语义的知识服务。关联课程数据的构建、组织以及知识管理是未来电子学习领域的重要研究方向。 本文从数据转换、关联课程数据构建、知识本体构建、关联数据存储索引等方面展开研究。 (1)在数据转换方面,提出了将多种类型的教学资源文档转换为RDF数据的方法,其中创新提出一种表格数据转换成关联数据的方法,本章提出一种表格文档转换成关联数据的方法,利用LOD数据集的语义关系,生成表中的列题、表元值、列与列之间关系的相关候选类和实体,然后进行关联推断,以马尔科夫网络图模型为框架,计算相应的因子节点的值,从候选集中选取最佳的类和实体,分配给表格中的列题和表元,最后将其转换成RDF数据,并和各经典数据集相关联。 (2)针对知识表示的问题,提出了关联课程数据构建方法,其中以计算机微机接口,组成原理等课程为例,构建计算机硬件课程RDF数据集。然后引入了知识本体的概念,介绍了本体的定义及本体的构建方法,知识本体的元数据以及知识点的认知顺序,在此基础上构建前序后续知识点之间的关联,是系统实现知识发现和知识导航的基础,同时提供了更为复杂的检索功能。由于自然语言和概念表述的多样性,知识点之间存在非前序后续关系或非映射的关联,本文提出一种新的基于内置文本匹配的关联方法,将自然语言中相互关联,却无法用机器进行推断的相关知识点,存放在文本匹配表中,通过行表元的信息读取,构建它们之间的关联,为关联课程数据增添更多语义联接,有助于学习者更全面的掌握课程领域知识,提高学习效率。 (3)针对关联数据存储和索引的问题,分析目前采用的主要的语义数据存储方式,分析其利弊,提出一个基于垂直划分、多索引及属性表的混合存储结构,提高存储和查询功能。 同时,基于权威数据集的实验证明了本文算法研究的有效性,实验结果表明本文对于课程关联数据的转换、存储方面所做的深入研究和提出的实现方案是具有创新性,并且是可行和有效的。
[Abstract]:The concept of relational data was introduced in 2006. Publishing data information with associated data technology is the most important step to implement the data World wide Web. The resource description framework published by W3C organization and the Web ontology language formalize the concept and the relationship between the concepts that appear in the document, and the relationship between the concepts in the document is formally defined by the resource description framework published by the W3C organization and the Web ontology language. The network data has the characteristics of global universality and machine readability. The most important role of associated data is association creation and data consolidation. At the same time, the related data technology also provides semantic knowledge service to the e-learning system. The construction, organization and knowledge management of related curriculum data are important research directions in the field of e-learning in the future. In this paper, data conversion, association curriculum data construction, knowledge ontology construction, relational data storage index and other aspects are studied. The main contents are as follows: (1) in the aspect of data conversion, this paper proposes a method to convert many kinds of teaching resource documents into RDF data, in which a new method of transforming tabular data into related data is proposed. In this chapter, a method of transforming tabular documents into associated data is proposed. By using the semantic relationship of LOD data set, the column title, table element value, candidate classes and entities related to the relationship between columns and columns are generated, and then the correlation inference is made. The Markov network graph model is used as the frame to calculate the value of the corresponding factor node, select the best class and entity from the candidate set, assign it to the column title and table element in the table, and convert it into RDF data, and associate it with each classical data set. (2) aiming at the problem of knowledge representation, this paper puts forward the method of constructing the data of related courses, in which the RDF data set of computer hardware course is constructed by taking the courses of computer interface, composition principle and so on as examples. Then it introduces the concept of knowledge ontology, introduces the definition of ontology and the construction method of ontology, the metadata of knowledge ontology and the cognitive order of knowledge points, on the basis of which, it constructs the relationship between pre-order and follow-up knowledge points. It is the basis of the system to realize knowledge discovery and knowledge navigation, and provides more complex retrieval function at the same time. Due to the diversity of natural language and conceptual representation, there is a non-preorder or non-mapping relationship between knowledge points. This paper proposes a new correlation method based on built-in text matching, which associates natural language with each other. But the related knowledge points that can not be inferred by machine are stored in the text matching table. By reading the information of row table elements, the association between them is constructed, and more semantic joins are added to the related curriculum data. It is helpful for learners to master the knowledge of curriculum field more comprehensively and improve the learning efficiency. (3) aiming at the problem of relational data storage and index, this paper analyzes the main semantic data storage methods adopted at present, analyzes its advantages and disadvantages, and proposes a hybrid storage structure based on vertical partition, multi-index and attribute table, which is based on vertical partition, multi-index and attribute table. Improve storage and query functions. At the same time, the experiment based on authoritative data sets proves the validity of this algorithm. The experimental results show that the in-depth research and implementation of curriculum related data conversion and storage is innovative. And it is feasible and effective.
【学位授予单位】：武汉大学
【学位级别】：博士
【学位授予年份】：2013
【分类号】：TP391.1;TP333
[Abstract]:The concept of relational data was introduced in 2006. Publishing data information with associated data technology is the most important step to implement the data World wide Web. The resource description framework published by W3C organization and the Web ontology language formalize the concept and the relationship between the concepts that appear in the document, and the relationship between the concepts in the document is formally defined by the resource description framework published by the W3C organization and the Web ontology language. The network data has the characteristics of global universality and machine readability. The most important role of associated data is association creation and data consolidation. At the same time, the related data technology also provides semantic knowledge service to the e-learning system. The construction, organization and knowledge management of related curriculum data are important research directions in the field of e-learning in the future. In this paper, data conversion, association curriculum data construction, knowledge ontology construction, relational data storage index and other aspects are studied. The main contents are as follows: (1) in the aspect of data conversion, this paper proposes a method to convert many kinds of teaching resource documents into RDF data, in which a new method of transforming tabular data into related data is proposed. In this chapter, a method of transforming tabular documents into associated data is proposed. By using the semantic relationship of LOD data set, the column title, table element value, candidate classes and entities related to the relationship between columns and columns are generated, and then the correlation inference is made. The Markov network graph model is used as the frame to calculate the value of the corresponding factor node, select the best class and entity from the candidate set, assign it to the column title and table element in the table, and convert it into RDF data, and associate it with each classical data set. (2) aiming at the problem of knowledge representation, this paper puts forward the method of constructing the data of related courses, in which the RDF data set of computer hardware course is constructed by taking the courses of computer interface, composition principle and so on as examples. Then it introduces the concept of knowledge ontology, introduces the definition of ontology and the construction method of ontology, the metadata of knowledge ontology and the cognitive order of knowledge points, on the basis of which, it constructs the relationship between pre-order and follow-up knowledge points. It is the basis of the system to realize knowledge discovery and knowledge navigation, and provides more complex retrieval function at the same time. Due to the diversity of natural language and conceptual representation, there is a non-preorder or non-mapping relationship between knowledge points. This paper proposes a new correlation method based on built-in text matching, which associates natural language with each other. But the related knowledge points that can not be inferred by machine are stored in the text matching table. By reading the information of row table elements, the association between them is constructed, and more semantic joins are added to the related curriculum data. It is helpful for learners to master the knowledge of curriculum field more comprehensively and improve the learning efficiency. (3) aiming at the problem of relational data storage and index, this paper analyzes the main semantic data storage methods adopted at present, analyzes its advantages and disadvantages, and proposes a hybrid storage structure based on vertical partition, multi-index and attribute table, which is based on vertical partition, multi-index and attribute table. Improve storage and query functions. At the same time, the experiment based on authoritative data sets proves the validity of this algorithm. The experimental results show that the in-depth research and implementation of curriculum related data conversion and storage is innovative. And it is feasible and effective.
【学位授予单位】：武汉大学
【学位级别】：博士
【学位授予年份】：2013
【分类号】：TP391.1;TP333
【相似文献】
相关期刊论文 前10条
1 陈定权;卢玉红;;图书馆在关联数据运动中的角色解析[J];图书馆建设;2014年03期
2 曹月珍;马建玲;;关联数据在图书馆的最新发展[J];图书馆学研究;2014年14期
3 黄永文;;关联数据在图书馆中的应用研究综述[J];现代图书情报技术;2010年05期
4 刘炜;;关联数据:概念、技术及应用展望[J];大学图书馆学报;2011年02期
5 李琳;;关联数据在图书馆界的应用与挑战[J];图书与情报;2011年04期
6 曾琦;;基于关联数据的图书馆云服务平台设计[J];图书馆学研究;2012年05期
7 王薇;欧石燕;;关联数据在图书馆领域的应用研究[J];新世纪图书馆;2012年09期
8 吴e，
本文编号：2450590
本文链接：https://www.wllwen.com/kejilunwen/jisuanjikexuelunwen/2450590.html
