Application of SVR in Quantitative Analysis of Wines
LUO Tao1, WEI Ji-ping2, ZHAO Yu-ping3, ZHANG Jun4
1. School of Computer Science and Technology,Tianjin Key Laboratory of Cognitive Computing and Application, Tianjin University, Tianjin 300072,China 2. School of Chemical Engineering and Technology, Tianjin University,Tianjin 300072,China 3. School of Chemical and Biological Science, Yantai University,Yantai 264005, China 4. Sino-French Joint-Venture Dynasty Winery LTD., Tianjin 300402, China
Abstract:Fourier transform infrared spectroscopy has been widely used in some related fields, thus induces the rapid development of quantitative analysis method based on Lambert-Beer’s Law and chemometrics in recent years. The selection of appropriate pre-processing method and calibration model is extremely crucial to the quantitative analysis. The present paper selected 30 wine samples and used infrared spectroscopy combined with vector regression algorithm SVR quantitative analysis model with standard normal variate, baseline correction and outliers elimination to analyze four representative components of wine. Satisfactory results were gained while the relative errors were limited to less than 5%. This method can be applied to the wine representative quantitative analysis for the material content.
[1] Gauglitz G, Vo-Dinh T. Handbook of Spectroscopy, Wiley, VCH, 2003. [2] Griffiths P R, Haseth J. Fourier Transform Infrared Spectrometry, Wiley, 2007. [3] Wold H. Soft Modeling by Latent Variables: the Nonlinear Literative Partial Least Squares Approach, Perspectives in Probability and Statistics, 1975. [4] Despagne F, Massart D L. Neural Networks in Multivariate Calibration, Analyst, 1998, 123: 157R. [5] Vapnik V. Statistical Learning Theory, Wiley, New York, 1998. [6] Belousov A I, Verzakov S A, von Frese J. Applicational Aspects of Support Vector Machines, J. Chemom., 2002, 16: 482. [7] Capron X, Smeyers-Verbeke J. Food Chemistry,2007,101(4):1585. [8] Ovidiu Lvanciuc. Reviews in Computational Chemistry, 2007, 23:291.