非相关文献知识发现的融合评价方法构建研究  

Research on the Development of the Fusion Evaluation Method for the Disjoint Literature-based Knowledge Discovery

作　　者：祝清松[1,2] 冷伏海[1] 

机构地区：[1]中国科学院国家科学图书馆,北京100190 [2]中国科学院大学,北京100049

出　　处：《情报理论与实践》2013年第7期106-109,105,共5页Information Studies:Theory & Application

摘　　要：评价对于非相关文献知识发现的发展具有重要促进作用。通过评价可以发现不同系统的优势和不足,引导非相关文献知识发现的研究方向。但是,目前的评价方法局限于某个系统,不能进行多个系统之间的比较。文章在对各种评价方法分析的基础上,认为当前建立一套标准的评价方法具有很大挑战性,依据融合评价原则,提出一个以统计评价方法为主、其他方法为辅的多方法融合评价策略。Evaluation plays an important role in promoting the development of disjoint literature-based knowledge discovery. By evaluation, we may find the superiorities and deficiencies of different systems, and lead the research directions of disjoint litera- ture-based knowledge discovery. However, the current evaluatioh methods are limited to a certain system, and can not be used to make comparisons among several systems. Based on the analysis of the various evaluation methods, this article argues that establis- hing a standardized set of evaluation methods is very challenging at present. According to the fusion evaluation rules, this article proposes a multi-method fusion evaluation strategy led by the statistical evaluation method and supplemented by the other methods.

关 键 词：非相关文献 知识发现 评价方法 

分 类 号：G350[文化科学—情报学]