自回归差分移动平均模型季节乘积模型在儿童EB/CMV/Cox病毒感染门急诊人次预测中的应用

Application of ARIMA seasonal product model to predict the number of children patients with EBV/COX/CMV infection

目的探讨自回归差分移动平均(autoregressive integrated moving average model,ARIMA)季节乘积模型在预测儿童EB、柯萨奇、巨细胞病毒三项(以下简称病毒三项)感染门急诊人次的应用。方法选择并收录本院2016年1月-2018年12月的儿科门急诊就诊并进行病毒三项检测的患儿人次。使用SPSS 17.0软件进行ARIMA季节乘积模型进行拟合,并计算2019年1月-4月的预测值与实际人次进行比较,评价模型预测效果。结果ARIMA(0,1,1)(0,1,0)12模型是拟合儿童病毒三项就诊人次的最佳预测模型,评价误差为20.46%。结论ARIMA季节乘积模型能为医院合理调配病毒三项患儿就诊资源提供有效依据。

Objective To investigate the clinical value of autoregressive integrated moving average(ARIMA)seasonal product model in forecasting the number of children patients with EBV/COX/CMV infection.Methods The outpatient and emergency department visits of pediatric patients who underwent three viral tests in our hospital from January 2016 to December 2018 were selected and included.SPSS 17.0 software was used to fit the ARIMA seasonal product model,and the predicted value from January to April 2019 was calculated and compared with the actual person-time,so as to evaluate the prediction effect of the model.Results ARIMA(0,1,1)(0,1,0)12 model was the best prediction model fitting the medical visit number of children patients with EBV/COX/CMV infection,with an evaluation error of 20.46%.Conclusion ARIMA seasonal product model can provide references to medical resources for patients with suspected infection of EBV/COX/CMV.