摘要
应用SIMCA分类法与PLS算法结合卷烟纸的傅里叶变换近红外光谱(FT-NIR)建立了卷烟纸的分类模型,用于卷烟纸的判别分类,效果良好;同时,建立了测定卷烟纸定量、厚度、透气度、水分和灰分等性质的校正模型,其相应的相关系数分别为0.976 8,0.966 4,0.947 0,0.956 3和0.975 9;全交互校验均方残差分别为0.561 4,0.096 0,1.274 1,0.096 7和0.260 3。校正模型应用于样品实测,结果准确,令人满意。
By using algorithm of SIMCA and partial least squares（PLS） combined with Fourier transform near infrared spectra （FT-NIR）, the classification methods were established for the discrimination of cigarette paper. Meanwhile, the calibration models were established for the determination of the gramrnage, thickness, permeability, moisture and ash of cigarette paper. Correlation coefficients of the models were 0.9768, 0.9664, 0. 9470, 0.9563 and 0. 9759, and the root mean square errors of cross validation （RMSECV） were 0. 561 4, 0.0960, 1.274 1, 0.0967 and 0. 2603 respectively. The methods has been applied to the determination and discrimination of unknown samples with satisfactory results.
出处
《光谱学与光谱分析》
SCIE
EI
CAS
CSCD
北大核心
2006年第10期1858-1862,共5页
Spectroscopy and Spectral Analysis
基金
云南省高教厅教学科研学术带头人项目资助
