利用THz时域谱技术和支持向量机回归法快速分析食用油成分

Rapid Analysis of Edible Oil Compositions by Terahertz Time-Domain Spectroscopy and Support Vector Regression Method

采用THz时域谱技术获得30个食用油样本的吸收谱,随机分为训练集和预测集。利用支持向量回归算法和气相色谱方法得到的参照测量值,在训练集上建立食用油中主要成分的预测模型,并用此模型对预测集样本进行预测。结果表明该模型对食用油预测的相关系数均在0.9以上,比传统PLS方法更为精确。

The terahertz band absorption spectrum of edible oil in 30 samples was obtained by terahertz time-domain spectroscopy（THz-TDS） technology.With support vector regression algorithm and reference concentration of oil compositions measured by gas chromatography,samples were randomly divided into training set and testing set to build a regression model on main components to predict the composition concentration.The results shows that the models achieve the correlation coefficient of 0.9 or more in actual-predicted relations,having significant advantage compared with the traditional method of PLS.