基于文本挖掘的领域本体半自动构建方法研究─以教学设计学科领域本体建设为例

Semi─automatic Building Approach of Domain Ontology Based on Text Mining：A Case Study of Building Instructional Design Domain Ontology

目前知识工程研究已逐渐成为教育技术学研究领域的热点问题，其中构建领域本体是知识工程的一个重要方面。从已有的基于本体的应用研究来看，领域本体可以在学习者模型建立、网络教学支撑平台开发、智能答拟系统开发、资源库建设等诸多方面发挥重要作用，而构建领域本体是各种应用得以实现的前提。然而构建领域本体是一项浩大的工程，需要大量领域专家的参与，当前是手工完成的，建设周期长。面对一个知识迅速增长的时代，仅仅依靠领域专家来手工构建领域本体显然不能满足需求。因此，本研究提出一种基于文本挖崛的领域本体半自动构建方法，即借助于面向文本的智能信息处理技术从教科书或科技资料中提炼概念、概念定义和概念之关的关系，并辅之以学科专家和知识工程师的参与，从而加快领域本体建设速度。整个构建过程由”文献材料选取”、”文献材料文本处理”、”概念关系标注”、”形式化表示与存储”、”知识库评价”等五个环节组成，涉及信息提取、中文分词、术语自动提取、术语定义自动抽取等技术以及基于术语部件的概念间属种关系、并列关系半自动发现方法和基于术互信息值的概念关联关系半自动发展方法等方法。文章还针对当前开发的全国中小学教师教育技术能力培训的学习资源需求，选取”教学设计”这一课程作为研究个案，以基于文本挖掘的领域本体半自动构建方法来建设教学设计学科领域本体，以便进一歩丰富和优化教师教育技术能力远程培训的学习资源。

Knowledge engineering has become the research hotspots of Educational Technology and building domain ontology is an important part of knowledge engineering. Based on the application studies based on ontology， domain ontology can play an important roel in the fields of building the model of learner， developing Learning Management System（LMS）， developing Intelligent Answering System， constructing resource database， etc.， and building domain ontology is the base of the above applications. At present， building domain ontology is a great project in which lots of domain experts join. The experts often manually build domain ontology and spend a long time. Facing with the time of knowledge quickly increasing， building ontology manually only by experts can’t meet the requirements. So in this study an semi─automatic building approach of domain ontology based on ext mining is brought forward， which is to extract concepts， concept definition and relationships between concepts from textbooks or scientific materials with text─oriented intelligent information processing technology and with the help of subject experts and knowledge engineers. This approach is constituted of the selection of document materials， the text processing of document materials， the concepts relation marking， the formal representation and storage and the evaluation of knowledge database， including the technologies of information extraction， Chinese word segmentation， segmentation， automatic terms extraction and automatic term definition extraction. At last， according to the learning resource requirement in Teachers Training of educational technology capability， the authors took the curriculum of Instructional Design as study case and built the domain ontology of instructional design with semi─automatic building approach of domain ontology based on text mining.