您的位置:首页  > 论文页面

生物医学文献中抽取药物间交互反应信息的实验研究

发表时间:2013-08-31  浏览量:1041  下载量:472
全部作者: 赵江,兰曼,时红垒
作者单位: 华东师范大学信息科学技术学院
摘 要: 第一届药物间交互作用信息抽取竞赛(drug-drug interaction extraction, DDIExtraction)是在生物信息学研究领域中,基于药物之间相互反应关系的抽取竞赛,是第一次尝试抽取药物之间交互作用(drug-drug interaction, DDI)的研究,为评价不同方法和系统提供了一个统一的平台。在该竞赛的数据集上进行实验,分析并比较了从文献中产生不同层面训练数据策略的优劣和评估不同特征类型的有效性。实验结果显示:1) 在药物实体层面上进行相互作用检测比在句子层面上进行更有效;2) 利用简单自然语言处理(natural language processing, NLP)的输出信息不能提高分类效果,需要使用更加高级的处理技术提高分类效果。
关 键 词: 计算机应用;药物间交互作用;对比实验;信息抽取;自然语言处理
Title: An exploration of drug-drug interaction extraction using biomedical texts
Author: ZHAO Jiang, LAN Man, SHI Honglei
Organization: School of Information Science and Technology, East China Normal University
Abstract: The first challenge of drug-drug interaction extraction (DDIExtraction) is to determine whether a pair of drugs appearing in a given sentence has interaction information. This challenge is the first attempt at extraction of drug-drug interaction (DDI) information in wide community, which provides an consolidate platform to evaluate different information extraction techniques. In this paper, based on this challenge, we compared and evaluated the effectiveness of different strategies of example generation from texts and different feature types for drug relation extraction. Our comparative results showed that 1) drug-drug interaction classification at drug entity pair level outperformed that at sentence level; 2) simple natural language processing (NLP) output did not improve performance and more advanced way of incorporating NLP output needed to be explored.
Key words: computer application; drug-drug interaction; comparative test; information extraction; natural language processing
发表期数: 2013年8月第16期
引用格式: 赵江,兰曼,时红垒. 生物医学文献中抽取药物间交互反应信息的实验研究[J]. 中国科技论文在线精品论文,2013,6(16):1501-1507.
 
0 评论数 0
暂无评论
友情链接