摘要
采用造纸法再造烟叶的近红外光谱,结合支持向量机建立了产品理指标包括定量、厚度、填充值、抗张强度及热水可溶物定量模型,评价了模型预测效果。近红外预测值与各物理指标之间具有很好的相关性,各模型均具有很好的预测准确性。定量、厚度、抗张强度、填充值及热水可溶物模型的RMSEP分别为1.156 7,0.006 2,0.028 7,0.106 9及0.803 3。通过对再造烟叶物理指标模型的验证表明该方法用于再造烟叶物理指标快速测定是可行的。
Near infrared spectroscopy of 1 000 paper-making reconstituted tobacco products were adopted, combined with support vector Regression (SVR), the physical indexes of reconstituted tobacco products were established including quantitative model of gramme, thickness, filling value, tensile strength and hot water soluble substances, and at the same time, the effect of model prediction was evaluated. There was a good correlation between the near infrared prediction value and the physical indexes, and the model had good prediction accuracy. The RMSEPs of the gramme, thickness, tensile strength, filling value and hot water soluble model were 1.156 7, 0.006 2, 0.028 7, 0.106 9 and 0.803 3, respectively. It was proved that the method could be used to determine the physical parameters of reconstituted tobacco products quickly.
出处
《食品工业》
北大核心
2017年第8期262-264,共3页
The Food Industry
基金
云南省科技厅项目"云产卷烟再造烟叶提质提量技术研究及应用"(2015IA026)
云南中烟科技项目"造纸法再造烟叶外加纤维的筛选与应用研究"(2015YL02)
关键词
近红外光谱:支持向量回归
造纸法再造烟叶
物理指标
near infrared spectroscopy
support vector regression
papermaking reconstituted tobacco
physical index