基于灰色模型和ARMA模型的猪瘟月新发生次数预测比较

Comparison of grey model and ARMA model for predicting the number of monthly new outbreaks of CSF

对猪瘟月新发生次数(每个月的新发生次数)的定量预测可以为当地动物疫病防控部门提供依据,使其针对猪瘟发病风险有选择性地采取防控措施。该文分别选用灰色预测模型和ARMA预测模型对贵州省猪瘟月新发生次数进行了预测。灰色模型和ARMA模型以2005-2008年兽医公报中统计的贵州省猪瘟发病数据为预测依据,以2009年的统计数据评价比较预测效果。灰色模型和ARMA模型预测平均绝对误差分别为1.84和1.48,平均绝对百分误差为0.272和0.229。预测结果表明,ARMA预测模型的预测精度更高,对贵州省猪瘟发病的预测是可行有效的。

The quantitative prediction of the number of monthly new outbreaks of CSF can provide animal disease prevention and control departments the basis for taking selective preventive measures. Grey model and ARMA model were chosen to predict the number of monthly new outbreaks of CSF in Guizhou province, China. By grey model and ARMA model, outbreaks of CSF in Guizhou province were predicted according to the statistical data from veterinary bulletin from 2005 to 2008, and both models were assessed by the statistical data in 2009. Mean absolute error of grey model and ARMA model was 1.84 and 1.48 respectively, while mean absolute percentage error of two models was 0.272 and 0.229. The results showed that the prediction precision of ARMA model was higher than grey model, and ARMA model was feasible and effective for the prediction of CSF outbreaks in Guizhou province, China.