摘要
目的探讨支持向量机对ARIMA算法的改进作用,同时利用某院的门诊诊疗人次对ARIMA-SVM组合模型进行实证研究,为SVM在医院的利用提供实践基础。方法从理论层面解释SVM算法以及ARIMA-SVM组合模型;再选取某院2014年1月-2016年12月的数据用ARIMA-SVM组合模型构建预测模型,并且预测2017年1月-2017年9月门诊诊疗人次,对比单纯ARIMA模型的预测结果,比较2个模型的预测能力。结果 ARIMA-SVM组合模型的预测精度优于单纯的ARIMA模型,ARIMA-SVM组合模型的MAPE为4.61%,而ARIMA模型的MAPE为4.90%。结论 ARIMA-SVM模型在医院运营管理中有积极作用,并且SVM为医院多项业务可提供支持。
Objectives To explore the improvement effect of support vector machine（SVM） on ARIMA algorithm. At the same time,we made an empirical study on ARIMA-SVM combination model by using outpatient visits in a hospital,which provideed a practical basis for the utilization of SVM in hospital. Methods First,from the theoretical level to explain what was the SVM algorithm,and ARIMA-SVM combination model; prediction model was constructed using ARIMA-SVM combination model then selected a hospital from January 2014 to December 2016 and January 2017 to September 2017 data,prediction of outpatient visits,prediction results of simple ARIMA model,prediction ability comparison of two models. Results The prediction accuracy of ARIMA-SVM combined model was better than that of pure ARIMA model,and the MAPE of ARIMA-SVM combination model was 4.61%,while the MAPE of ARIMA model was 4.90%. Conclusions ARIMA-SVM model played a positive role in hospital operation management,and SVM provideed support for multiple operations in hospital.
作者
杜军
郭慧敏
黄路非
Du Jun;Guo Huimin;Huang Lufei(The Third People＇s Hospital of Chengdu in Sichuan Province,Chengdu 610031,China)
出处
《中国病案》
2018年第9期40-42,共3页
Chinese Medical Record
