广东省结核病发病趋势的时间序列分析

目的 利用乘积季节模型预测广东省结核病的发病趋势，为制订结核病的防控措施提供科学的依据。 方法 利用最小二乘法原理，应用自回归求和移动平均模型与随机季节模型相结合的乘积季节模型，对广东省结核病发病趋势进行预测。 结果利用1996年至2008年资料构建ARIMA（0,1,1）（0,1,1）₄模型，所建立的模型的预测效果良好，实际值均在预测值的95%可信区间内。 结论 采用ARIMA乘积季节模型预测广东省结核病发病情况，拟合及预测效果较好，预测结果符合全省发病现状及当前采取的防控措施。

Application of time series analysis for the prediction of incidence trend of tuberculosis in Guangdong province

Objective To utilize the product seasonal model to predict the trend of TB and provide scientific evidence for formulating the related measures of prevention and cure. Methods The product seasonal model was combined with ARIMA model and stochastic seasonal model by using the least square principle. Results The ARIMA（0,1,1）（0,1,1）4 model was established by the data from 1996 to 2008, and the effectiveness of prediction of this model showed to be good with the actual values in the 95% confidence interval of predicted values. Conclusion The ARIMA product seasonal model shows effective to predict the incidence of TB in Guangdong province, and the results is in according to the current status of TB, moreover, it could provide information for us to take measures for TB prevention and control.