摘要
为了解决传统领域知识的主题图构建方法中大量时间和人力耗费的问题,提出了面向课程领域知识的主题图学习方法,分析了TMLDK的主要活动流程,建立了学习活动图,讨论了TMLDK中领域关键词提取、关键词向主题的转换以及信息集成等关键技术,并利用浅层解析、相似度计算等技术完成了网络课件向TMC的半自动转换,最后利用TM4J和Java开发出TMLDK的原型,对其学习过程进行了验证。TMLDK的实现能够协助领域专家方便地构建领域知识的主题图,可以节省大量的人力和时间,而且有利于TMC后期的维护和更新。
In order to solve the problem of much time costing in traditional topic map constructing for domain knowledge, a method oftopic map learning for domain knowledge （TMLDK） is put forward, and the main steps of TMLDK is analyzed as well as its activityflow is built. And then the key technology has been discussed on domain keywords extracting, conversion from the keyword to topicand information integration. Afterthat, the courseware is converted to TMC half-automatically through the technology of shallow parsingand similarity computing. Finally the prototype of TMLDK is developed by the use TM4J and Java, which verifies the model. TMLDKis able to help the expert to build topic map for the domain knowledge, which can save plenty of manpower and time, and moreover, isconvenient for upper maintenance and update of TMC.
出处
《计算机工程与设计》
CSCD
北大核心
2007年第18期4529-4531,共3页
Computer Engineering and Design
