摘要
以拉曼、近红外2种光谱特征融合结合化学计量学方法对花生油掺伪进行了定量分析。分别用激光拉曼、激光近红外光谱仪采集134个掺伪油样本的光谱数据,采用SPXY算法对样本集进行划分。拉曼光谱(Ram)和近红外光谱(near infrared spectroscopy,NIR)数据进行预处理后,采用后向间隔偏最小二乘法(BiPLS)和联合间隔偏最小二乘法(synergy interval partial least squares,Si PLS)分别提取2种光谱的特征波长;将提取的特征波长融合,结合支持向量机回归(SVR)建立数学模型,采用网格搜索算法(CV)对SVR模型的参数组合(C,g)值寻优,建立最优参数模型。研究表明:建立的Ram-NIR-SVR模型能够实现花生油中掺杂油脂含量的快速准确预测,预测集和校正集的相关系数R分别达到0.98和0.99,均方根误差(MSE)低于2.38E-3;对比不同特征波长提取方法,并与单光谱分析技术比较,可以看出,数据融合技术能够增强模型预测能力,减小模型参数,有利于模型的实际应用,体现了2种光谱很好的互补性。表明光谱分析结合数据融合技术对食用油真实性综合鉴别具有重要意义。
The purpose of this study is to conduct quantitative analysis on the adulteration in peanut oil by combining data fusion of Raman and near infrared（ NIR） spectral characteristics with chemometrics methods. With laser Raman and NIR spectrometer,the spectra of 134 adulterated oil samples were collected. The sample set partitioning algorithm based on joint X-Y distances（ SPXY） was employed to divide the samples. The spectra data of Raman and NIR were preprocessed. Backward interval partial least squares（ Bi PLS） and synergy interval partial least squares（ SiPLS） were used to extract the characteristic wavelengths of the spectra data. On the intervals of data fusion of characteristic wavelengths of two kinds of the spectra,adulteration quantity prediction models were established by Support Vector Machine Regression（ SVR）. In the end,the author optimized the combination of model parameters（ C,g） by Mesh Search Algorithm and determined the optimal parameter model. According to the analysis,the model which was established by SVR based on combined Raman and near-infrared（ NIR） spectral data could implement the content prediction of the adulteration oil content of peanut oil. Furthermore,the correlation coefficient R of prediction set and calibration set can reach 0. 98 and 0. 99 respectively,and the root mean square error（ MSE） was smaller than 2. 38E-3. Compared with single spectral analysis and different characteristic extraction methods,the results showed prediction ability was enhanced and the parameter was reduced by using the data fusion technology. Practical application of the model is favorable. And it reflects the good complementarity of Raman and near infrared spectrum. So it is significant to study the authenticity identification of edible oils by combining spectral analysis and data fusion technology.
出处
《食品与发酵工业》
CAS
CSCD
北大核心
2016年第4期169-173,共5页
Food and Fermentation Industries
基金
国家"十一五"科技支撑计划项目(2009BADB9B08)
武汉市科技攻关计划项目(2013010501010147)
武汉工业学院食品营养与安全重大项目培育专项(2011Z06)
武汉轻工大学研究生创新基金项目(2014cx005)
关键词
花生油
拉曼光谱
近红外光谱
定量分析
数据融合
支持向量机回归
peanut oil
Raman spectroscopy
near infrared spectroscopy（NIR）
quantitative analysis
data fusion
support vector machine regression（SVR）
