热带病与寄生虫学

热带病与寄生虫学 ›› 2015, Vol. 13 ›› Issue (3): 156-158,186.doi: 10.3969/j.issn.1672-2302.2015.03.010

• 论著 • 上一篇    下一篇

ARIMA模型在预测医疗机构收治重症手足口病患者趋势中的应用

王燕，陈萍，刘红，段晓菲，陈岚，林华   

  1. 610000 成都市，成都市公共卫生临床医疗中心
  • 出版日期:2015-09-10 发布日期:2015-09-30

Application of ARIMA model to predicting severe hand-foot-mouth disease trend in medical institutions

Wang Yan, Chen Ping, Liu Hong, Duan Xiaofei, Chen Lan, Lin Hua   

  1. Chengdu Public Health Clinical Medical Center, Chengdu 610000, China.
  • Online:2015-09-10 Published:2015-09-30

PDF (PC)

1176

摘要： 【摘要】 目的  应用自回归积分滑动平均模型（ARIMA）分析预测医疗机构收治重症手足口病趋势，为医疗机构合理调配人、财、物力提供依据。 方法  应用SPSS 16.0软件对2010~2013年逐月收治重症手足口病情况进行拟合，用2014年就诊数据验证所得的ARIMA模型，并对2015年就诊趋势进行预测。 结果  通过对参数和模型的拟合优度检验以及残差白噪声序列的检验，最终确定模型为ARIMA(0，1，0)(1，1，0)12，其标准化BIC=6.346、平稳R2=0.708、均方根误差=21.576，LB统计量检验残差序列为白噪声序列。 结论  ARIMA模型可较好预测医院收治的重症手足口病趋势，为医院管理提供依据，模型预测效果的优化有待于原始数据和其他相关数据的持续积累。

关键词: 自回归积分滑动平均模型, 手足口病, 重症患者, 预测, 时间序列分析

Abstract:

【Abstract】 Objective  To assess the significance of applying autoregressive integrated moving average(ARIMA) model to predicting the trend of severe hand-foot-mouth disease(HFMD) for evidences to rationally allocate human, financial and material resources in a medical institution. Methods  The monthly severe HFMD cases obtained from 2010 to 2013 were fitted in SPSS(version 16.0), and the data collected in 2014 were used to verify the previously established ARIMA model by which it was used to predict the prevalence trend of HFMD in 2015. Results  The model ARIMA(0,1,0)(1,1,0)12 were established based on the test of parameters and goodness of fit as well as sequence of white-noise residuals, in which the normalized BIC was defined as 6.346, stationary R square as 0.708, and root mean square error as 1.576. LB test for residuals sequence as defined as white noise sequence. Conclusion  ARIMA model can well predict the prevalence trend of severe HFMD, and may supply evidences for medical institutions in management of this disease. However, the effectiveness of this model remains yet to be optimized on continuous accumulation of the original data and other related data basis.

Key words: Autoregressive integrated moving average, Hand-foot-mouth disease(HFMD), Severe patient, Prediction, Time series analysis