时间序列分解法在预防接种门诊接种量预测的应用  

Application of a time series decomposition method for predicting vaccine doses administered in immunization clinics

作　　者：黄美林[1] 刘世科[1] 胡丹标[1] 王帆[1] 章群英 

机构地区：[1]宁海县疾病预防控制中心,浙江宁海315600 [2]宁海县城关医院,浙江宁海315600

出　　处：《中国疫苗和免疫》2017年第6期681-684,共4页Chinese Journal of Vaccines and Immunization

基　　金：浙江省预防医学会软科学研究课题项目(编号2015YF03);宁海科技局社会发展类科技计划项目(编号201618)

摘　　要：目的探讨适用预防接种门诊疫苗接种量的预测模型。方法收集宁海县城关预防接种门诊2010-2014年各月常规免疫剂次数,应用时间序列分解法建立接种剂次数的趋势值(Y)随时间(t)的回归方程,预测2015年各月接种剂次数。结果该门诊月接种剂次数的趋势值回归方程为Y=6 750.383 8-15.289 5 t,据此预测2015年该门诊各月接种剂次数与实际接种剂次数的相对误差在0.45%-25.44%之间。结论应用时间序列分解法对预防接种门诊接种量的预测具有良好的可行性。Objective To develop a model to predict doses of vaccine administered in immunization clinics. Methods We collected counts of doses administered by month in Chengguan immunization clinic of Ninghai county from 2010 to 2014,and used a time series decomposition method to build a regression equation with doses administered（ Y） and time（ t） to predict doses administered by month for the following year-2015. Results The resulting regression equation was Y = 6 750. 383 8-15. 289 5 t. The relative error between predicted administered doses by month in the immunization clinic in 2015 ranged from0. 45% to 25. 44%,compared with actual doses administered values. Conclusions It is feasible to use a time series decomposition method to predict administered vaccine doses in immunization clinics.

关 键 词：时间序列分解法 预测模型 接种剂次 

分 类 号：R186[医药卫生—流行病学]