基于烟叶物理属性的烟叶出片率模型研究

为研究烟叶物理属性与出片率的定量关系,应用逐步回归、主成分回归、支持向量机回归3种方法对58个烟叶样品建立出片率预测模型.同时,采用3种预测模型对15个未知烟叶样本进行预测.结果显示:逐步回归模型预测标准偏差为1.168;主成分回归预测标准偏差为1.203;当参数c=29.744,g=0.01时,支持向量机预测标准偏差为0.624.表明采用支持向量机回归方法预测精度最高.

Study on Leaf Yield Model of Tobacco Based on Physical Properties of Tobacco Leaves

In order to study the quantitative relationship between physical attributes and leaf yield of tobacco leaves,stepwise regression,principal component regression and support vector machine regression were used to establish a prediction model for leaf yield of 58 tobacco leaves.At the same time,three prediction models were used to predict 15 unknown tobacco leaf samples.The results show that the standard deviation of stepwise regression model is 1.168;that of principal component regression is 1.203;and that of support vector machine is 0.624 when the parameter c=29.744 and g=0.01.The results show that the prediction accuracy of support vector machine regression method is the highest.