基于深度学习的医疗问答系统的开发  

Development of Medical Question-and-Answer System Based on Deep Learning

作　　者：姚智[1] YAO Zhi(Department of Information,Nantong First People’s Hospital,Nantong Jiangsu 226001,China)

机构地区：[1]江苏省南通市第一人民医院信息科

出　　处：《中国医疗设备》2019年第12期88-91,141,共5页China Medical Devices

摘　　要：目的为了促进我国医疗卫生行业与新兴技术的深度融合,推动我国医疗信息化事业的发展,本文设计并实现了一种基于深度学习的医疗问答系统。方法综合采取检索式与生成式两种方法,使用网络爬虫技术以及人工审核的方式构建了医疗知识库,使用Seq2Seq算法构建了答案生成模型。结果所开发的医疗问答系统能够与用户进行有效交流并提供相关的医疗建议,并且通过专家审核的方式保证了其准确性。结论基于深度学习的医疗问答系统实现了深度学习技术与医疗行业的结合,推动了医疗信息化的发展。Objective To promote the deep integration of medical and health industry and emerging technologies, and promote the development of medical informationization industry, a medical inquiry system based on deep learning was designed and implemented. Methods The search and generation methods were comprehensively adopted, the medical knowledge base was constructed by using web crawling technology and manual review, and the answer generation model was constructed by using the Seq2Seq algorithm. Results The developed medical question and answer system was able to communicate effectively with users and provide relevant medical advice, and its accuracy was guaranteed by expert review. Conclusion The medical question-andanswer system based on deep learning reflects the combination of deep learning technology and medical industry, and promotes the development of medical informationization.

关 键 词：医疗信息化 深度学习 Seq2Seq序列 问答系统 深度长短期记忆 

分 类 号：TP311.5[自动化与计算机技术—计算机软件与理论] TP277[自动化与计算机技术—计算机科学与技术]