论文部分内容阅读
目的分析潍坊市农药中毒报告的变化趋势,探讨整合自回归移动平均模型(ARIMA)在农药中毒报告数拟合中的应用。方法利用潍坊市2006-2014年农药中毒报告资料,采用SPSS 19.0专家建模器建立最优预测模型,预测潍坊市2015年农药中毒报告数。结果 2006-2014年间潍坊市非生产性农药中毒报告数自2011年起略有下降趋势。应用ARIMA(0,0,0)(1,0,1)模型、简单季节模型较好地拟合了既往时间序列的生产性与非生产性农药中毒报告数,生产性模型的R2、正态化BIC分别为0.413、3.594,非生产性模型的R2、正态化BIC分别为0.727、5.604。结论模型能较好地描述潍坊市2006-2014年农药中毒报告数的分布规律,预测效果较好。
Objective To analyze the trend of pesticide poisoning report in Weifang and to explore the application of ARIMA in the fitting of pesticide poisoning report number. Methods Based on the report of pesticide poisoning in Weifang City from 2006 to 2014, the optimal forecasting model was established by SPSS 19.0 expert modeller to predict the 2015 pesticide poisoning reports in Weifang City. Results The number of reports of unproductive pesticide poisoning in Weifang from 2006 to 2014 decreased slightly from 2011 onwards. Using the ARIMA (0,0,0) (1,0,1) model, the simple seasonal model well matches the past time series of productive and unproductive pesticide poisoning reports, the R2 of the productive model, normality BIC were 0.413, 3.594, R2 for unproductive models, and normalized BICs were 0.727 and 5.604, respectively. Conclusion The model can well describe the distribution of pesticide poisoning reports in Weifang City from 2006 to 2014, and the prediction result is better.