摘要
用近红外光谱数据建立了相思树综纤维素含量的预测模型。首先对近红外光谱数据进行预处理并按波长进行分组,然后使用拟合方法建立许多非线性子数学模型,最后通过计算加权平均值给出预测综纤维素含量的公式。模型预测值的平均相对误差为0.0074,预测误差明显小于用近红外光谱仪所带软件建立的模型的预测误差。文中的建模方法有望用于相思树其它成分的预测。
A model for predicting the contents of acacia holocellulose（acaciarichii1905A6Fv） was established by using the Near Infrared Spectrometry（NIR）.First,the near infrared spectrum data were preprocessed and were divided into several groups according to their wavelengths.Then,many nonlinear sub-mathematical models were established by using a fitting method.Finally,the equation for predicting the contents of acacia holocellulose was given by calculating the weighted mean value.The model had its mean relative prediction error of 0.0074 which was obviously less than that of the model established by using the specified software for the near infrared spectrometer.This modelling method was possible to be used to predict the other contents of avacia.
出处
《红外》
CAS
2010年第5期37-40,共4页
Infrared
基金
国家科技支撑计划项目(2006BAD32B03-5)
关键词
近红外光谱
数学模型
综纤维素
相思树
near infrared spectrometry
mathematical model
holocellulose
acacia
