目的建立时间序列模型,分析应用模型预测新型冠状病毒肺炎流行初期日累计病例数的效果,为更好地防控新冠肺炎疫情提供参考。方法采用时间序列方法建立新型冠状病毒肺炎日累计病例数的时间序列模型,通过模型拟合预测疾病未来短期日累计病例数,通过比较从模型得到的日累计病例数预测值与实际值、分析中位绝对误差对模型的预测效能进行评估分析。结果时间序列模型对湖北地区、非湖北地区日累计病例数的预测结果与实际发展趋势基本一致,中位绝对误差分别为6.40%、2.23%,对非湖北地区日累计病例数的预测效能优于湖北地区。结论采用时间序列模型对新型冠状病毒肺炎流行初期的日累计病例数进行预测的效果较好,但不能真实地反映出疾病的发展变化趋势,现实中必须考虑其他因素对预测结果的影响。
ObjectiveTo establish a time series model and to analyze the effect of applying the model to predict the daily cumulative cases of the COVID-19 epidemic in the early stage, so as to provide a reference for better COVID-19 prevention and control. MethodsThe time series method was used to establish a time series model for daily cumulative cases of COVID-19. By means of the model, the daily cumulative cases of the disease were predicted in the short-term. According to the comparison between the predicted value of the daily cumulative cases from the model and actual value, the median absolute deviation was analyzed to evaluate the prediction efficiency of the model. ResultsThe prediction results of the time series model for the daily cumulative cases in Hubei and non-Hubei areas were basically consistent with the actual development trend. The median absolute deviation in Hubei and non-Hubei areas was 6.40% and 2.23%, respectively. The prediction efficiency of the daily cumulative cases in non-Hubei areas was superior to that in Hubei areas. ConclusionThe effect of using the time series model to predict daily cumulative cases of COVID-19 epidemic in the early stage is relatively good, but it cannot truly reflect the development and change of the disease. The impact of other factors for the prediction results must be considered in reality.