季节性指数平滑法在我国肺结核发病数预测中的应用

Prediction of tuberculosis incidence in China based on seasonal index smoothing method

目的预测我国结核病发病情况,为有效防治结核病提供参考。方法利用2015年1月至2018年12月我国肺结核发病例数数据,采用指数平滑法进行拟合预测。结果指数平滑法的Winters相乘模型为最优预测模型,平稳的R^(2)分别为0.859,R^(2)值为0.842;标准化的BIC为16.663,模型残差为白噪声序列(Ljung-Box Q为26.924,P>0.05);Winters相乘模型在Alpha(水平)、Gamma(趋势)、Delta(季节)均有统计学意义(P<0.05),该模型预测值与实际值的平均相对误差为2.57%。结论指数平滑法的Winters相乘模型能够较好地预测我国肺结核的发病情况,预测数据可为结核病的防治提供参考,值得进一步推广应用。

Objective To predict the incidence of tuberculosis in China and provide a reference for effective prevention and treatment of tuberculosis. Methods Using the data of the number of cases of tuberculosis in China from January 2015 to December 2018, the exponential smoothing method was used for fitting prediction.Results The Winters multiplication model of the exponential smoothing method was the optimal prediction model. The stationary R^(2) was 0.859, the R^(2) value was 0.842, the normalized BIC was 16.663, and the model residual was white noise sequence(Ljung-Box Q was 26.924, P>0.05), the Winters multiplication model was statistically significant(P<0.05) in Alpha(horizontal), Gamma(trend), and Delta(P<0.05). The average relative error between the predicted and actual values of the model was 2.57%.Conclusion The Winters multiplication model of exponential smoothing method can better predict the incidence of tuberculosis in China. The prediction data can provide reference for the prevention and treatment of tuberculosis, which is worthy of further promotion and application.