摘要
运用逐步回归周期分析方法，即均值生成函数、逐步回归分析方法的综合应用，对郑州冬小麦冬前正积温变化趋势进行预测。１９８９～１９９８年历史年代回代结果平均绝对误差２８．６℃，平均相对误差３．９％；１９９９、２０００年预报与实测值相当接近。
By way of stepwise periodic regression analysis, that is to say the synthetic using of mean value function and stepwise regression analysis, we forecast the change trend of fore - winder positive accumulated temperature for winder wheat in Zhengzhou. Form 1989 to 1998, the historic episode regressive answer s mean absolute error is 28.6℃, and its mean relative error is 3 .9%, In 1999 and 2000, the data gotten by forecast are very close to those which was observed.
出处
《河南气象》
2001年第4期33-34,共2页
Meteorology Journal of Henan
关键词
增值生成函数
逐步回归
冬前正积温
冬小麦
适播期
Mean value function
Stepwise regression
Fore - winter positive accumulated temperature
Forecast
