摘要
供应链对企业竞争力具有巨大意义,而供应链中最重要的部分就是企业供应关系,现有的公司实体关系抽取方法没有考虑关系中公司实体的角色,不适用于企业供应关系抽取。基于此,本文采用人工构建和自动构建相结合的方式构建了关系指示词库,利用关系指示词库判断文本的主题,并使用最近句法依赖动词对实体之间的语义关系进行判断。最后在上市公司年报文本上进行了测试,取得了不错的效果。
A good supply chain is indispensable for enterprises to improve competitiveness. For a supply chain, the most important part is the supply relationship between enterprises. Existing methods of extracting corporate entity relationship did not consider the role of corporate entity in the supply relationship. Therefore, these methods are not suitable for extracting enterprise supply relationship. To solve this problem, a library of relation word is constructed by combining manual construction with automatic construction. The relation word is used to judge the theme of the text, and the nearest syntactic dependent verbs are used to judge the semantic relationship between entities. Experiments on the annual report of listed companies have met the expected requirement.
出处
《计算机科学与应用》
2018年第12期1823-1832,共10页
Computer Science and Application
基金
国家自然科学基金资助项目(61502262).
