基于大规模真实文本的数值知识元挖掘研究

Numeric knowledge element mining based on large-scale realistic corpora

探讨了从海量文献中挖掘知识元的背景,并详述了从海量年鉴文本中抽取宏观数值知识元的基本流程和各主要环节的算法,并重点对数值知识元主体抽取的效果进行了分析,实验结果表明在特定领域内知识元挖掘要达到实用水平是可行的。

This paper discusses the necessity of the knowledge element mining,then describes in detail the base process and algorithms of main steps of extraction numeric knowledge elements from China yearbook full-text database,and then specially analyzes the extraction quality.