基于深度学习技术的便民服务创新模式研究

Research on the Innovation Mode of Convenient Service Based on Deep Learning Technology

目的:针对患者在就诊过程中缺乏专业指导带来的诸多问题,提出基于深度学习与就诊场景相结合的技术方法,使患者能够采用自然语义的方式描述疾病症状,获得系统推荐最佳科室和专业的咨询服务,从而促进便民服务模式创新。方法:采用医疗知识图谱和卷积神经网络方法,构建智能分诊模型,实现疾病知识的自动问答和智能导诊。基于面向服务(Service-Oriented Architecture,SOA)的开放架构,通过共享服务模式向传统预约服务渠道提供统一规范的服务接口。结果:在实际示范应用过程中,2018年智能分诊模型通过便民服务渠道提供了1246次/天的患者咨询,患者满意度达到91%。结论:基于深度学习技术的便民服务能够有效改善患者服务体验,提升患者就诊效率,降低人工服务成本,适合在医院推广使用。

Objective:In view of many problems caused by lack of professional guidance when patients seek medical service,this paper puts forward a technical method combining deep learning with medical visiting scene,which allows patients to describe disease symptoms in a natural semantic way to obtain suitable department recommended by the system as well as professional convenient consultation,thereby promoting the innovation of convenient service mode.Methods:By medical knowledge graph and convolution neural network,an intelligent triage model is built to realize automatic question answering of disease knowledge and intelligent triage.Based on the open-ended Service-Oriented Architecture(SOA),a unified and standardized service interface is provided to traditional reservation channels through the shared service mode.Results:In the practical demonstration application,the intelligent triage model provided patient consultation 1,246 times/day in 2018 through convenient service channels,and patient satisfaction reached 91%.Conclusion:The convenient service based on deep learning technology can effectively improve patients service experience,promote medical visiting efficiency and reduce labor service cost,which should be popularized in hospital.