基于PSO的LSTM神经网络门诊量预测研究  

Study on Outpatient Quantity Prediction of Long Short-term Memory Neural Network Based on Particle Swarm Optimization

作　　者：覃智威 宓轶群[2] 陆允敏[2] 刘钊 朱平[1] QIN Zhi-wei;MI Yi-qun;LU Yun-min;LIU Zhao;ZHU Ping(School of Mechanical Engineering,Shanghai Jiao Tong University,Shanghai 200240,China;Shanghai Sixth People’s Hospital,Shanghai 200233,China;School of Design,Shanghai Jiao Tong University,Shanghai 200240,China)

机构地区：[1]上海交通大学机械与动力工程学院,上海200240 [2]上海交通大学附属第六人民医院,上海200233 [3]上海交通大学设计学院,上海200240

出　　处：《软件导刊》2020年第10期29-33,共5页Software Guide

基　　金：上海市卫生和计划生育委员会项目(201740034);上海市卫生和计划生育委员会中医院科技创新项目(ZYKC201701006);徐汇区重要疾病联合攻关项目(XHLHGG201802);上海市医院协会医院管理研究基金项目(X1901072)。

摘　　要：研究基于粒子群优化算法的长短期记忆神经网络模型构建方法及其在医院门诊管理中的应用,选取三甲医院中医慢病相关科室门诊量历史记录数据,构建基于长短期记忆神经网络的门诊量预测模型,借助粒子群优化算法对长短期记忆网络进行参数优化,并使用优化后的模型对门诊量进行预测。月度门诊量预测结果表明,经过粒子群优化的长短期记忆神经网络模型在测试集上的预测误差RMSE,相比未优化的模型减小了48.5%。粒子群优化算法能高效地优化预测模型,可使模型较好地预测出门诊量变化趋势,从而为医务人员门诊管理工作提供决策支持。To investigate the construction method of long short-term memory(LSTM)neural network model based on particle swarm optimization(PSO)algorithm and its application in hospital outpatient management,the historical outpatient quantity data of the department related to Chinese medicine chronic disease in the tertiary hospital are selected to establish an outpatient quantity prediction model by means of the long short-term memory neural network,and particle swarm optimization algorithm is used to optimize the parameters of the prediction model.The prediction results of monthly outpatient quantity show that the prediction error RMSE of the LSTM neural network model optimized by PSO on the test set is 48.5%lower than that of the unoptimized model.Particle swarm optimization algorithm can optimize the prediction model efficiently,so that the model can better predict the trends of outpatient quantity and provide decision support for the outpatient management of medical staff.

关 键 词：门诊量预测 长短期记忆神经网络 粒子群优化 深度学习 

分 类 号：TP319[自动化与计算机技术—计算机软件与理论]