摘要目的 探索运用人工神经网络及贝叶斯决策理论建立神经系统疾病门诊智慧分诊决策树管理模型.方法 以贝叶斯决策理论为理论基础,以人工神经网络技术完成神经系统疾病快速专科/亚专科机器学习;针对神经系统疾病专科或亚专科分诊数据,以循环神经网络及贝叶斯算法完成神经系统疾病症状与诊断的概率分布及收敛,建立神经系统疾病决策树管理模型并完成理论论证.结果 完成了神经系统疾病智慧分诊的管理理论及模型构建,根据迁移学习特性基本实现神经系统疾病的快速学习和精确分诊.结论 该管理模型的研究,能为后续应用提供理论借鉴意义,并在一定程度上缓解目前患者退换号率较高的问题.

abstractsObjective To develop an effective decision tree management model for smart triage of nervous system diseases based on artificial neural networks and Bayesian decision theory. Methods Bayesian decision theory was used as the theoretical basis, and convolutional neural network was used to complete the rapid specialist / sub-specialist machine learning. For the specialist or sub-specialist triage data, circular neural network and Bayesian algorithm were performed to complete the probability distribution and convergence of disease symptoms and diagnosis. Results The decision tree management model and theoretical demonstration were established. According to the characteristics of the transfer learning, the rapid learning of nervous system diseases and accurate triage system, and the remote smart triage system were successfully constructed. Conclusions The management model could provide theoretical references for further use, and alleviate to some extent the currently high rate of outpatient appointment withdrawal and changes.