| 王 颖,张智雄,孙 辉,雷 枫.国史知识的语义揭示与组织方法研究[J].中国图书馆学报,2015,41(4):55~64 |
| A Method of Semantic Representation and Organization of the Historic Knowledge on Contemporary China |
| 国史知识的语义揭示与组织方法研究 |
| |
| DOI：10.13530/j.cnki.jlis.150020 |
| Key words:Chinese national history Ontology Text mining Knowledge organization |
| 中文关键词: 国史 本体 文本挖掘 知识组织 |
| 基金项目:本文系中国社会科学院哲学社会科学创新工程信息化项目“中华人民共和国史教育网”的研究成果之一 |
|
| Hits: 3524 |
| Download times: 1328 |
| Abstract: |
China has a huge volume of historic resources on its contemporary history. However the organization of those historic resources is not satisfactory. Lots of valuable knowledge are hidden in those resources and can not be discovered and utilized easily. It is an urgent problem to mine the implicit semantic knowledge scattered in a large number of historic resources and to organize the knowledge based on its internal relations. The authors believe that it will benefit the education and research of contemporary China history to represent and organize those historic resources semanticly. Based on the historic encyclopedia of the contemporary China, the authors try to explore a new method to semantically represent and organize the historic knowledge hidden in the text resources, so as to help the users to represent the historic knowledge in a fine-grained manner, to reorganize the historic knowledge in a more semanticly enriched way, and to develop more innovative application by using the historic knowledge.
On the basis of related researches, this paper proposes a method, which is called “Mining down, Organizing up”, to represent and organize historic knowledge on contemporary China. Based on contemporary Chinese historic ontology, this method extracts knowledge objects and facts from the unstructured historic text items, forms a historic knowledge network of the contemporary China, and realizes the multidimensional knowledge organization on a higher level by the relations such as time, subclass, hierarchy, and statistics. Aiming at the key problems in the method, this paper puts forward specific procedures of representing and organizing historic knowledge of the contemporary China. Firstlly, the paper builds an ontology conceptual model as the basic organization mode dscribing the skeleton of historic knowledge of the contemporary China; Secondly, the paper determines core objects to populate ontology instance library served as corpus for automatic identification. After that, the paper extracts facts automatically from text items by unitilizing text mining technology, then ensures the reliability of these facts through judgment and complement of domain experts; Moreover, the paper builds historic knowledge network based on the association among knowledge objects; Finally, the paper realizes multidimensional knowledge organization on the basis of the network.
Based on this method, the authors represented and organized historic knowledge of the contemporary China from text resources, and developed a system to implement historic knowledge visualization, reorganization and other new applications including knowledge maps, relatedness analysis, and national history facts reconstruction and etc.
The limitations of this research are that 1) the accuracy of recognizing knowledge objects and relevant facts from text should be improved, especially the identification and recommendation of relevant national history facts; it will further reduce the workload of domain experts; 2) The association calculation method of historic knowledge network is simple and has not fully applied current semantic similarity calculation and graph mining methods. These are key problems that need to be solved in our future studies.
Studies show that the “Mining down, Organizing up” method can realize fine grained representation of historic knowledge of the contemporary China and innovative application of knowledge organization based on historic knowledge objects; it can be used as a kind of new knowledge representation and organization methods applicable in other fields.
This paper proposes an original method for semantic representation and organization, which uses contemporary Chinese historic ontology as basic organization mode, extracts important knowledge objects and facts from the historic text items with text mining technology, forms a complex network, and realizes multidimensional knowledge organization by assocation calculation on the network. The “Mining down, Organizing up” method effectively realized the fine grained representation and deep level organization of historic knowledge of the contemporary China. 7 figs. 12 refs.
|
| 中文摘要: |
| 深入开展中华人民共和国国史的教育和研究一直是各方高度重视的工作。将国史知识进行语义揭示和组织,对于国史的教育和研究具有重要意义。在参考相关研究成果的基础上,本文提出了“向下挖掘,向上组织”的国史知识语义揭示与组织方法。这一方法以国史本体为基础,在对隐藏于国史资源文本条目中的国史知识对象和相关事实进行语义挖掘和揭示的基础之上,通过国史知识对象的关联,构建国史知识网络,并基于时间、类属、层级及统计等关系,对国史知识内容进行更高层次的多维组织展示,并基于这一思路开发了相应的系统平台,实现了国史知识的揭示、重组和其他创新应用。本文所提出的国史知识语义揭示和组织方法对其他类型知识的开发利用具有参考价值。图7。参考文献12。 |
|
View Full Text
View/Add Comment Download reader |
|
|
|
