基于R语言ARIMA模型在慢阻肺急性加重患者发病预测中的应用

发布时间：2018-07-22 14:09

【摘要】：目的建立慢阻肺急性加重入院人次的自回归积分滑动平均模型(ARIMA),科学预测慢阻肺急性加重入院人次,为该病的诊治以及合理利用医疗资源提供理论依据。方法使用R语言(v.3.2.3)做模型的识别、模型的参数估计与检验,建立ARIMA模型,对某院2013-2015年慢阻肺急性加重出院人次进行模型拟合,用2016年1~3月的预测值与实际值作比较,检验模型的预测能力,并且预测2016年4-6月慢阻肺急性加重入院人次。结果经过多次检验,确定ARIMA(2,2,1)(1,1,1)_(12)模型预测能力最佳,其残差序列是白噪声。用2016年1~3月数据来检验模型,其MAPE的绝对值均小于10%,说明模型的拟合优度相对较好,预测能力可靠,根据该模型预测2016年4~6月该院慢阻肺急性加重入院人次分别为162、160、159。结论 ARIMA模型能够很好的拟合慢阻肺急性加重的入院人次并进行短期预测,模型显示2016年该院的急性支气管炎的入院人次将有所上升,为医院合理利用医疗资源提供了有力依据。
[Abstract]:Objective to establish an autoregressive integral moving average model (Arima) for predicting acute exacerbation of COPD in order to provide a theoretical basis for the diagnosis and treatment of the disease and the rational use of medical resources. Methods R language (v.3.2.3) was used to identify the model, the parameters of the model were estimated and tested, and Arima model was established. The model fitting was performed on the discharge times of acute exacerbation of COPD from 2013 to 2015 in a hospital. The predicted values from January to March 2016 were compared with the actual values. Test the predictive capacity of the model and predict acute exacerbation of COPD in April-June 2016. Results after many tests, the Arima (2H _ 2N _ 1) (1T _ 1) _ (12) model has the best prediction ability, and its residual sequence is white noise. Using the data from January to March 2016 to test the model, the absolute values of MAPE were all less than 10, which indicated that the model had relatively good fit and reliable prediction ability. According to the model, the number of acute exacerbations of COPD in the hospital from April to June 2016 was 162160159, respectively. Conclusion Arima model can well fit the admission number of patients with acute exacerbation of chronic obstructive pulmonary disease and make short-term prediction. The model shows that the admission of acute bronchitis in this hospital will increase in 2016. It provides a strong basis for the rational utilization of medical resources in hospitals.
【作者单位】：成都市第三人民医院信息部;
【分类号】：R563.9

【参考文献】