摘要
目的:向有健康咨询需求的患者推荐高专业度和高活跃度的医生,进一步完善在线问诊平台功能。方法:融合Word2Vec和层次分析法构建医生推荐模型,以“好大夫在线”网站作为实证对象,验证模型适用性。首先,基于患者特征,借助Word2Vec和余弦相似度,得到相似患者集合,通过医生和患者的实际问诊关系得到医生推荐序列C。然后,运用层次分析法建立医生诊断能力评分体系,计算出各项指标权重及医生得分,按照得分递减顺序输出医生推荐序列D。融合医生推荐序列C和医生推荐序列D得到最终医生推荐集合E。结果:该模型综合考虑了医生的专业程度和活跃程度,所推荐的医生在临床经验上具备较高的相似度,且在线活跃度较高。结论:融合Word2Vec和层次分析法的模型具有较好的推荐效果,实践过程中需充分调研患者需求以确保模型更具代表性。
Objective To recommend highly professional and active doctors to patients with health consultation needs,and further improve the functions of the online consultation platform.Methods A doctor recommendation model was constructed by combining Word2Vec and analytic hierarchy process(AHP),and www.haodf.com was taken as the research object to verify the model.Firstly,with the help of Word2Vec and cosine similarity,a set of similar patients was obtained based on the patient characteristics,and doctor recommendation sequence C was obtained through the actual consultation relationship between the doctor and the patient.Then,AHP was used to establish the doctor's diagnostic ability scoring system,calculate the weight of each index and the doctor's score,and output doctor recommendation sequence D in decreasing order.The final doctor recommendation set E was obtained by fusing doctor recommendation sequence C with doctor recommendation sequence D.Results This model comprehensively considered the doctors’levels of professionalism and activity,and the recommended doctors had high similarity in clinical experience and high online activity.Conclusion The model proposed in this paper combining Word2Vec with AHP has a good recommendation effect but it is necessary to fully investigate the needs of patients to ensure that the model is more representative.
作者
王妞妞
熊回香
刘樱
WANG Niu-niu;XIONG Hui-xiang;LIU Ying(School of Information Management,Central China Normal University,Wuhan 430070,Hubei Province,China;Department of Gastrointestinal Surgery Ⅰ Section,Renmin Hospital of Wuhan University,Wuhan 430070,Hubei Province,China)
出处
《中华医学图书情报杂志》
CAS
2022年第1期20-31,共12页
Chinese Journal of Medical Library and Information Science
基金
国家社会科学基金年度项目“融合知识图谱和深度学习的在线学术资源挖掘与推荐研究”(19BTQ005)。
