健康社区用户教育信息推荐模型比较

目的:探讨不同推荐模型对基于健康社区的高血压用户推荐教育信息的效果，为慢性病患者健康教育服务模式提供依据。方法:搜集权威可信的高血压患者健康教育文章以及健康社区用户提问，构建并比较潜在狄利克雷分配（LDA）推荐模型、语义推荐模型及向量空间模型的效果，为用户提问提供精准信息。结果:LDA推荐模型的平均准确度均值最高，向量空间模型次之，语义推荐模型最低。结论:在虚拟健康社区用户问答过程中，LDA推荐模型可提升健康教育信息推荐精度，实现用户个性化推荐。

Comparison of information recommendation models for education of health community consumers

Objective To study the effect of different recommendation models on recommended education information for hypertensive health community consumers in order to provide the basis for health education service model of chronic disease patients. Methods The authoritative and reliable papers on health education of hypertensive patients and questions of health community consumers were collected, latent Dirichlet allocation (LDA) recommendation model, semantic recommendation model and vector space model were established and their effects on recommended education information for hypertensive health community consumers were compared in order provide the accurate and precise information for them. Results The mean accurate and precise rate of LDA recommendation model was the highest followed by vector space model and semantic recommendation model. Conclusion LDA recommendation model can increase the accuracy and precision of health education information recommendation in questions and answers of virtual health community consumers.