基于动态集成加权概率RF的门诊量预测  

Outpatient volume prediction based on dynamic integrated weighted probability RF

作　　者：樊冲 FAN Chong(Jinzhou Big Data Management Center,Jinzhou 121000,Liaoning,China)

机构地区：[1]锦州市大数据管理中心,辽宁锦州121000

出　　处：《智能计算机与应用》2024年第5期209-214,共6页Intelligent Computer and Applications

摘　　要：医院门诊量本质上是一种具有潜在规律的时间序列,通过对门诊量进行有效分析和预测,可以更加科学、合理地配置医疗资源。针对门诊量波动幅度较大的时间序列预测问题,提出一种基于动态集成加权概率RF的门诊量预测方法。首先选择具有强泛化性的随机森林(Random Forest,RF)作为预测模型;并且采用k近邻-层次聚类算法对RF模型中树的强度进行评估,从中动态选择性能最佳的决策树,提高回归模型的性能;为了提升预测模型的准确率,采用加权概率融合规则代替原始RF模型的求平均数的规则。经过与BP神经网络和RF对比实验结果表明,提出方法可以更加精准地对门诊量进行预测和分析,为医院更好的运营管理提供了重要依据和决策支持。Hospital outpatient volume is essentially a time series with potential laws.Through effective analysis and prediction of outpatient volume,medical resources can be allocated more scientifically and reasonably.Aiming at the time series forecasting problem of large fluctuation of outpatient volume,a forecasting method of outpatient volume based on dynamic integrated weighted probability RF is proposed.Firstly,Random Forest(RF)with strong generalization is selected as the prediction model.Furthermore,K nearest neighbor-hierarchical clustering algorithm is used to evaluate the strength of trees in RF model,and the decision tree with the best performance is dynamically selected to improve the performance of regression model.Then,in order to improve the accuracy of the prediction model,the weighted probability fusion rule is used instead of the average rule of the original RF model.Compared with BP neural network and RF,the proposed method can predict and analyze outpatient volume more accurately,which provides the important basis and decision support for better hospital operation and management.

关 键 词：门诊量 随机森林 k近邻-层次聚类 加权概率融合 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]