应用乘积季节模型预测狂犬病的发病趋势

Feasibility for application of season model to predict incidence of rabies

目的探讨应用乘积季节模型模型进行狂犬病发病率预测的可行性,为降低狂犬病发病率提供理论依据。方法收集全国2004-2009年狂犬病发病率数据。采用差分方法对序列资料进行平稳化,进行定阶,建立2010年全国狂犬病发病率预测模型,并对预测结果进行分析和评价。结果乘积季节模型拟合结果较理想,很好地拟合了既往时间段上的发病率序列,预测2010年各月发病率较2009年有所下降。结论用乘积季节模型对狂犬病发病率数据拟合较为满意,预测效果良好,可为进一步制定预防策略措施提供依据。

The aim of the present study was to explore the feasibility for application of season model to predict the rabies incidence in China so as to provide the theoretical basis for continuing to reduce the rabies incidence.season model was established based on the rabies incidence of China from 2004 to 2009.Using difference method to smooth the sequence,the order was determined and the 2010 national rabies incidence forecast model was established to evaluate the predicting results.The results indicated that season model fitted well and the residual autocorrelation function graph showed the residuals were white noise sequences.The prediction results showed that rabies incidence and the predicted month-incidence by the model was consistent with the actual incidence.The fitting result in season model of the incidence of the rabies incidence is satisfactory,and the forecasting achieves good effects,which also provides scientific basis for the prevention and control of rabies incidence.