摘要
[目的/意义]旨在为企业建立自己的专家信息库提供参考。[方法/过程]构建一个统一的专家信息元数据框架,对专家信息进行了规范化描述;对专家主页的网页标题、网页URL和网页内容等进行了分析,并从中提取特征进行专家主页的识别;构建一个自适应多种主页类型的Bi-LSTM-CRF专家信息抽取模型,并对模型性能进行了实验分析。[结果/结论]基于专家主页的专家信息抽取方法能够利用机器学习技术从海量的Web页面中有效识别专家主页,并且通过改进的信息抽取模型从专家主页中自动抽取企业所需要的专家信息。本文提出的方法是有效的。
[Purpose/significance]The paper is to provide references for enterprises to establish their own expert information base.[Method/process]The paper builds a unified expert information metadata framework,describes the expert information in a standardized way;analyzes the title,URL and content of the expert homepage to extract features to identify the expert homepage;builds an adaptive expert information extraction model for various homepage types,and analyzes the model performance.[Result/conclusion]The expert information extraction method based on expert homepage proposed in this paper can use machine learning technology to effectively identify the expert homepage from a large number of web pages,and automatically extract expert information from the expert homepage through an improved information extraction model.The experimental results show that the proposed method is effective.
作者
曹洪飞
顾复
张今
陈芨熙
Cao Hongfei;Gu Fu;Zhang Jin;Chen Jixi(Department of Industry and System Engineering Zhejiang University,Hangzhou Zhejiang 310027)
出处
《情报探索》
2019年第12期1-9,共9页
Information Research
基金
国家重点研发计划课题“科技资源分享模型与开放分享理论”(项目编号:2017YFB1400302)研究成果之一
关键词
专家资源
元数据框架
专家主页识别
专家信息抽取
expert resources
metadata framework
expert homepage recognition
expert information extraction
