Extracting Higher Order Relations From Biomedical Text
Extracting Higher Order Relations From Biomedical Text
复制标题
从生物医学文本中提取高阶关系
DOI：
10.3115/v1/w14-2114
复制
发表时间：
2014
期刊：
影响因子：
--
通讯作者：
Robert E. Mercer
中科院分区：
文献类型：
--
作者：
Syeed Ibn Faiz;Robert E. Mercer
科学文章中的论证是由新旧知识的未表达和明确的陈述组成的，这些陈述组合成逻辑上连贯的文本论证。语篇关系是连接语篇各部分的语言连贯关系，有助于沟通论证的逻辑步骤。生物医学关系展示生物医学实体之间的关系。在本文中，我们主要关注的是提取生物医学关系之间的连接，我们称之为高阶关系。我们结合联合收割机两种方法，即生物医学关系抽取和话语关系解析，提取这种高阶关系的生物医学研究论文。发现和提取这些关系可以帮助自动理解作者在文本中表达的科学论点。
Argumentation in a scientific article is composed of unexpressed and explicit statements of old and new knowledge combined into a logically coherent textual argument. Discourse relations, linguistic coherence relations that connect discourse segments, help to communicate an argument’s logical steps. A biomedical relation exhibits a relationship between biomedical entities. In this paper, we are primarily concerned with the extraction of connections between biomedical relations, a connection that we call a higher order relation. We combine two methods, namely biomedical relation extraction and discourse relation parsing, to extract such higher order relations from biomedical research articles. Finding and extracting these relations can assist in automatically understanding the scientific arguments expressed by the author in the text.