基于自然语言处理技术的定题监测功能实现研究  

Design and Implementation of Automatic Monitoring Function Based on Natural Language Processing Technology

作　　者：刘巍[1] 王思丽[1] 祝忠明[1] 吴志强[1] 

机构地区：[1]中国科学院兰州文献情报中心,甘肃兰州730030

出　　处：《图书与情报》2018年第3期135-140,共6页Library & Information

基　　金：中国科学院西部之光青年项目"基于学术大数据的专题化信息自动采集与组织技术研究"(项目编号:Y6AX021001)研究成果之一

摘　　要：文章主要描述了在自动监测功能研发过程中,如何引入自然语言处理相关技术,从而提高开放知识资源自动监测采集过程的准确性、通用性、可配置性及松耦合性。研究发现,通过将自然语言处理技术应用在自动监测功能中,可以实现对监测资源中重要概念和实体的自动抽取,并与经过用户配置的语料库进行相似度匹配,最终基于匹配的结果实现自动化定题监测的目标。实践应用证明,文章提出的基于自然语言处理技术的定题监测方法目前已应用在相关项目的建设中且实测效果较好,证明其在一定程度上改进了传统的定源定向监测采集方法,提高了监测结果的准确性,优化和简化了监测参数的配置流程,有效提升了功能的通用性和松耦合性。This paper describes how to apply natural language processing technology in the development of automatic monitoring functions,improving the accuracy,versatility,configurability and loose coupling of the process of automatic monitoring and acquisition of open knowledge resources.The application of the natural language processing technology can extract important keywords and entities and similarity match with configuration item which configured by users.Finally,based on the matching results,system can determine whether the target is focused.so as to achieve the goal of automated monitoring.This method has been applied in the development of IIBD platform and has a positive effect.This study has improved the traditional fixed-source monitoring method.The accuracy of monitoring results was improved,and configuration of monitoring parameters were optimized and simplified,and versatility and loose coupling of functions were increased.

关 键 词：自然语言处理 实体抽取 相似度计算 定题监测 信息采集 

分 类 号：TP312[自动化与计算机技术—计算机软件与理论]