摘要
目的探讨应用指数平滑法预测黑热病未来发病数的可行性。方法应用SPSS13.0对2003～2008年甘肃省黑热病分季度发病数建立指数平滑模型,预测2009年分季度发病数,并用预测值与实际值进行比较,检验预测效果。结果该序列存在线性趋势和季节趋势,建立指数平滑模型,其常规参数α=0.1,季节参数δ=0.8,趋势性衰减π=0.1,平均相对误差为12.7%。结论指数平滑模型较好地拟合并预测了黑热病既往发病周期性以及各季度发病数,将其应用于甘肃省黑热病发病预测是可行的。
Objective To discuss feasibility of exponential smoothing in kala-azar incidence prediction in Gansu Province.Methods SPSS13.0 software was used to construct the exponential smoothing model based on the quarterly kala-azar incidence in Gansu Province from 2003 to 2008,to predict quarterly incidence in 2009,and kala-azar actual incidence in 2009 was compared with the incidence predicted by exponential smoothing model.Results The sequence had linear trend and seasonal trends,the exponential smoothing model was constructed with conventional parameters α = 0.1,seasonal parameter δ= 0.8,the trend of decay π = 0.1 and mean relative error of 12.7%.Conclusions The model exactly fits the incidence of the previous quarterly incidence and predicts the future incidence trend,thus,the exponential smoothing model for predicting kala-azar incidence in Gansu Province is feasible.
出处
《地方病通报》
2010年第3期24-25,29,共3页
Endemic Diseases Bulletin
关键词
指数平滑模型
预测
黑热病
Exponential smoothing model
Prediction
Kala-azar
