基于逐步回归的黑龙江省烟叶产量预测

东北优质烤烟生产决策系统开发与应用，主旨是为了开发一个系统，使用与烤烟产量相关因素的统计数据建立模型嵌入系统，开发出图形界面，使得农户或者烟草公司只要输入与烟叶产量相关因素的数据，后台通过模型计算就可以立即得到烟叶产量的预测数据。因此应用逐步回归的理论，利用MATLAB数学软件的数理统计工具箱，选取了与烟叶产量相关的移栽时间、5-8月平均气温以及施肥等因素，建立黑龙江省烟叶产量与这些相关因素间的逐步回归模型，找到对烟叶产量有显著影响的因素，结果表明：5-8月平均气温、移栽时间、纯磷肥施用量、纯钾肥施用量以及饼肥施用量对烟叶产量有显著影响，根据得到的结论，建议技术人员重视移栽时间对烟叶产量的影响，并且进一步规范烟农在生产实践中各种肥料的使用量。

The Prediction of Tobacco Yield of Heilongjiang Province Based on Stepwise Regression

Northeast quality flue-cured tobacco production and decision-making system development and application, it is to develop a system, then to product statistical data modeling embedded systems of using related factors of tobacco yield, and develop a graphical interface, as long as farmers or tobacco companies make the input and yield data related to factors, finally calculate by the model can be immediately cured tobacco production data. So this paper applied the theory of stepwise regression, and made use of the toolbox of mathematical statistics of Matlab software, selected the related factors with tobacco yield, such as transplanting time, average temperature from May to August and fertilization, then established stepwise regression equation with tobacco yield of Heilongjiang Province and these factors. Found out the factors that had a significant effect to tobacco production. The results showed that there was a significant effect to tobacco yield that average temperature from May to August, transplanting time, the number of pure phosphorus fertilizer per acre applied, the number of pure K fertilizer per acre applied, the number of pure P cake fertilizer per acre applied. According to the conclusions, to suggest technical person to pay attention to the influence of transplanting time to tobacco yield, and further regulate the number of fertilizer usage in the production.