烟草2种病害的增长模型及赤星病预测模型研究

Growth Model of Two Tobacco Diseases and Forecast Model of Tobacco Brown Spot Disease

用SPSS21.0软件中的二次函数等6种不同数学模型对烟草赤星病和靶斑病病情指数随时序值而变化的数据进行分析,通过决定系数R2、标准误和F值对模型进行比较。结果表明:逻辑斯蒂模型能够较好地模拟烟草赤星病和靶斑病的病情发展,三次函数也能较好地模拟烟草赤星病的病情发展。以8月中旬烟草赤星病病情指数为预测对象,温度、湿度、降雨量和日照时数为预测因子,用SPSS21.0软件建立多元回归预测模型。通过相关性分析筛选出7个关键预测因子,建立了多元回归预测模型:Y=-42.654-0.285X_(1)+4.26X_(2)-0.361X_(3)+0.162X_(4)+0.04X_(5)+0.103X_(6)-0.015X_(7)。通过逐步多元回归分析进行最优选择,得到多元回归预测模型:Y=-35.63+2.286X_(2)+0.157X_(4)。经拟合度检验,该模型的平均拟合率为86.49%。该模型的建立对于烟草赤星病的病害预测及防治提供了理论基础。

Six different mathematical models in SPSS21.0 software,such as quadratic function,were used to analyze the data of disease index of tobacco brown spot disease and target spot disease changing with time.These models were compared by determining coefficient R2,standard error and F value.The results showed that logistic model can well simulate the development dynamics of tobacco brown spot disease and target spot disease,and cubic function can also better simulate the development dynamics of tobacco brown spot disease.Taking the disease index of tobacco brown spot disease in mid-August as the forecast object,temperature,humidity,rainfall and sunshine hours as the forecast factors,SPSS21.0 software was used to establish multiple regression forecast model.Seven key predictors were selected by correlation analysis,and a multiple regression forecast model was established:Y=-42.654-0.285X_(1)+4.26X_(2)-0.361X_(3)+0.162X_(4)+0.04X_(5)+0.103X_(6)-0.015X_(7).The optimal selection was conducted by stepwise multiple regression analysis,and the multiple regression model was obtained:Y=-35.63+2.286X_(2)+0.157X_(4).The average fitting rate of the model is 86.49%through the test of fitting degree.The establishment of the model provides a theoretical basis for tobacco brown spot disease prediction and control.