摘要
【目的/意义】为了帮助科研人员从海量信息中发现热点和重大研究进展,抽取出有用的事件信息。【方法/过程】采用条件随机场方法和语义角色标注技术,构建了模型进行训练和学习。【结果/结论】提出了科技事件抽取框架,实现了科技事件抽取系统,取得了一定的抽取效果,该系统的可扩展性和可移植性有待提高。
【Purpose/significance】To help researchers discover hot spots and the major findings from the mass information and extract useful event information.【Method/process】The paper uses conditional random fields method and semantic roles labelling, and builds model to train and learn.【Result/conclusion】Put forward the scientific event extraction framework,realize it and achieve a certain effect.The scalability and portability of the system need to be further improved.
作者
刘振
LIU Zhen(Xuzhou Institute of Technology,Xuzhou 221008,China;University of Chinese Academy of Sciences,Beijing 100190 China;University of Chinese Academy of Sciences,Beijing 100190,China)
出处
《情报科学》
CSSCI
北大核心
2018年第9期115-117,122,共4页
Information Science
基金
国家社会科学基金(09FTQ005)
国家十二五科技支撑计划项目(2011BAH10B00)
关键词
科技事件
事件抽取
抽取框架
触发词
scientific event
event extraction
extraction architecture
trigger
