摘要
目的探讨时间序列分析法在手足口病预测中的应用,分别构建基于百度指数与发病率的自回归积分滑动平均模型(ARIMA),对比模型预测效果从而选出较佳的模型。方法疫情资料来源于中国疾病预防控制信息系统,百度关键词指数数据从百度指数官网上以月为单位进行收集。采用stata15.0SE与SPSS 20.0软件对数据进行整理与分析,并建立ARIMA模型;运用均方根百分比误差(RMSPE)和平均绝对百分比误差(MAPE)来评价百度指数构建的ARIMA模型与发病率构建的ARIMA模型之间的预测效果。结果2017-2021年深圳市辖区手足口病共9174例,年均发病率为16.43/10万;发病率高峰集中在10、5、7月。男性年均发病率10.33/10万,高于女性年均发病率6.11/10万,差异有统计学意义(χ^(2)=16.215,P<0.05)。发病年龄段主要集中在0~<6岁(7960例,86.76%),人群种类分布以散居儿童为主(6901例,75.22%)。基于复核百度指数的模型MAPE=41.535%,RMSPE=21.633%;基于发病率的模型MAPE=42.731%,RMSPE=22.201%,前者的预测性能更好。结论基于百度指数建立的ARIMA模型可提升模型的预测性能,可结合手足口病发病率及百度指数建立ARIMA模型,加强预测效果。
Objective To explore the application of time series analysis in the hand-foot-and-mouth disease,construct autoregressive integrated moving average model(ARIMA)based on Baidu index and incidence,and compare the prediction effect of the models.Methods Stata15.0SE and SPSS 20.0 software were used to sort out and analyze the data,and an ARIMA was established.The root-mean-square percentage error(RMSPE)and mean absolute percentage error(MAPE)were used to evaluate the prediction effect between ARIMA model constructed by Baidu index and ARIMA model constructed by incidence.Results From 2017 to 2021,a total of 9174 hand-foot-and-mouth disease cases were reported in Shenzhen city,with an annual incidence rate of 16.43/100000.The peak season of the incidence concentrates in October,May and July.The average incidence rate of male was 10.33/100000,higher than that of female(6.11/100000);the difference was statistically significant(χ^(2)=16.215,P<0.05).The age group of the incidence concentrated mainly 0-<6 years old(7960 cases,86.76%);the population type distribution was mainly scattered children(6901 cases,75.22%).MAPE=41.535%and RMSPE=21.633%based on Baidu index;MAPE=42.731%and RMSPE=22.201%based on incidence model;the former prediction performance was better.Conclusion The ARIMA model based on Baidu index could improve the prediction performance of the model,and the ARIMA model could be established by combining the incidence of hand-foot-and-mouth disease and Baidu index to enhance the prediction effect.
作者
毕嘉琦
郑磊
梁玲莎
钟玉君
BI Jiaqi;ZHENG Lei;LIANG Lingsha;ZHONG Yujun(Public Health Service Center,Bao'an District of Shenzhen,Shenzhen,Guangdong 518104,China)
出处
《热带医学杂志》
CAS
2023年第9期1323-1327,共5页
Journal of Tropical Medicine
基金
深圳市宝安区基础研究(医疗卫生)项目(2021JD250)
