电信科学 ›› 2019, Vol. 35 ›› Issue (4): 139-145.doi: 10.11959/j.issn.1000-0801.2019078

• 运行商人工智能专栏 • 上一篇    下一篇

基于自然语言学习的智能云导诊技术

汤人杰1,江涛2(),杨巧节1   

  1. 1 中国移动通信集团浙江有限公司,浙江 杭州 310016
    2 浙江省卫生计生信息中心,浙江 杭州 310006
  • 修回日期:2019-04-10 出版日期:2019-04-20 发布日期:2019-04-25
  • 作者简介:汤人杰(1983- ),男,中国移动通信集团浙江有限公司高级工程师,主要研究方向为大数据、人工智能。|江涛(1977- ),男,浙江省卫生计生信息中心高级工程师、副主任,主要研究方向为医疗健康行业信息化。|杨巧节(1984- ),女,中国移动通信集团浙江有限公司工程师,主要研究方向为人工智能。

Intelligent cloud guidance technology based on natural language learning

Renjie TANG1,Tao JIANG2(),Qiaojie YANG1   

  1. 1 China Mobile Communications Group Zhejiang Co.,Ltd.,Hangzhou 310016,China
    2 Health and Family Planning Information Center of Zhejiang Province,Hangzhou 310006,China
  • Revised:2019-04-10 Online:2019-04-20 Published:2019-04-25

摘要:

分析了目前导诊系统存在的主要问题,根据目前学术界研究的智能导诊系统现状,创造性地提出了基于海量患者病情自述,利用多种机器学习算法,形成了病情自述自学习体系,实现了针对病人自述的智能导诊。该平台利用网络爬虫技术获取了国际国内主流医疗机构科室设置以及海量的病人病情描述,形成了基于国际标准的病理知识库及病情自述知识库,为后续的智能文本识别奠定了基础。同时在算法上进行了创新,构建了注意力(attention)与文本积卷(TextCNN)组合模型,提升了导诊的准确性。

关键词: 智能导诊, attention模型, TextCNN

Abstract:

The main problems existing in the current guidance system were analyzed.According to the intelligent guidance system researched in the current academic circles,a self-learning system for the disease which based on a large number of patient condition descriptions using a variety of machine learning algorithms was creatively proposed.The platform used Web crawler technology to obtain setting of the international and domestic medical institution departments and a large number of patient descriptions,which form a pathological knowledge base and a disease self-reported knowledge base.At the same time,the algorithm was innovatively constructed to combine the attention model with TextCNN to improve the accuracy of the consultation.

Key words: intelligent guidance, attention model, TextCNN

中图分类号: 

No Suggested Reading articles found!