Abstract：The application of least square support vector machines (LS-SVM) regression method based on statistics study theory to the analysis with near infrared (NIR) spectra of tomato juice was introduced in the present paper. In this method, LS-SVM was used for establishing model of spectral analysis, and was applied to predict the sugar contents (SC) and available acid (VA) in tomato juice samples. NIR transmission spectra of tomato juice were measured in the spectral range of 800-2 500 nm using InGaAs detector. The radial basis function (RBF) was adopted as a kernel function of LS-SVM. Sixty seven tomato juice samples were used as calibration set, and thirty three samples were used as validation set. The results of the method for sugar contents (SC) and available acid (VA) prediction were: a high correlation coefficient of 0.990 3 and 0.967 5, and a low root mean square error of prediction (RMSEP) of 0.005 6° Brix and 0.024 5, respectively. And compared to PLS and PCR methods, the performance of the LS-SVM method was better. The results indicated that it was possible to built statistic models to quantify some common components in tomato juice using near-infrared (NIR) spectroscopy and least square support vector machines (LS-SVM) regression method as a nonlinear multivariate calibration procedure, and LS-SVM could be a rapid and accurate method for juice components determination based on NIR spectra.
黄康，汪辉君，徐惠荣*,王剑平，应义斌 . 基于最小二乘支持向量机的番茄汁糖酸度分析研究[J]. 光谱学与光谱分析, 2009, 29(04): 931-934.
HUANG Kang, WANG Hui-jun, XU Hui-rong*,WANG Jian-ping, YING Yi-bin . NIR Spectroscopy Based on Least Square Support Vector Machines for Quality Prediction of Tomato Juice. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2009, 29(04): 931-934.
[1] FU Xia-ping, YING Yi-bin, LU Hui-shan, et al(傅霞萍, 应义斌, 陆辉山, 等). Spectroscopy and Spectral Analysis(光谱学与光谱分析), 2007, 27(5): 911. [2] CHEN Nian-yi, LU Wen-cong, YE Chen-zhou, et al(陈念贻, 陆文聪, 叶辰洲, 等). Computers and Applied Chemistry(计算机与化学应用), 2002, 19(6): 691. [3] Vapnik V N. The Nature of Statistical Learning, 2nd. ed. New York：Springer, 2000. [4] Belousov A I, Verzakov S A, Von Frese J. Chemometrics and Intelligent Laboratory Systems, 2002, 64: 15. [5] ZHANG Lu-da, SU Shi-guang, WANG Lai-sheng, et al(张录达, 苏时光, 王来生, 等). Spectroscopy and Spectral Analysis(光谱学与光谱分析), 2005, 25(1): 33. [6] ZHANG Lu-da, JIN Ze-chen, SHEN Xiao-nan, et al(张录达, 金泽宸, 沈晓南, 等). Spectroscopy and Spectral Analysis(光谱学与光谱分析), 2005, 25(9): 1400. [7] BAI Peng, XIE Wen-jun, LIU Jun-hua(白 鹏, 谢文俊, 刘君华). Spectroscopy and Spectral Analysis(光谱学与光谱分析), 2007, 27(7): 1323. [8] Suykens J A K, Gestel T V, de Brabanter J, et al. Least Squares Support Vector Machines. Singapore: World Scientific, 2002. [9] YAO Xiao-gang, DAI Lian-kui(姚肖刚, 戴连奎). Proceeding of the 5<sup>th</sup> World Congress on Intelligent Control and Automation, June 15, 2004, Hangzhou，China(第五届全球智能控制与自动化大会2004年6月15日，中国杭州). [10] Chauchard F, Cogdill R, Roussel S, et al. Chemometrics and Intelligent Laboratory Systems, 2004, 71: 141. [11] Alessandra B, Marco F F, Cesar M, et al. Analytica Chimica Acta, 2006, 579: 25. [12] GUO Hui, LIU He-ping, WANG Ling(郭 辉, 刘贺平, 王 玲). Journal of System Simulation(系统仿真学报), 2006, 18(7): 2033. [13] Curda L, Kukacková O. Journal of Food Engineering, 2004, 61(4): 557. [14] Bülent U. A Comparison of Support Vector Machines and Partial Least Squares Regression on Spectral data, 2003. [15] Thissen U, Pepers M, Ustun B, et al. Chemometrics and Intelligent Laboratory Systems, 2004, 73: 169.
