A Feasibility Study on the Use of Visible Spectroscopy to Classify Different Brands of Vinegar

摘要

The feasibility of using visible spectroscopy to classify different brands of vinegar was analyzed in this research. Four brands of vinegars (Shanxi Chencu; Zilin; Donghu and Ninghuafu) and a total of 240 samples were prepared for the discrimination analysis. Orthogonal Signal Correction (OSC) was applied in this research for it can remove information from the spectral data that is not necessary for fitting of the concentration information by orthogonal processing. Partial Least Squares (PLS) analysis was implemented to build the discrimination model. The result shows that the correlation coefficient of calibration and validation model is 0.977 and 0.962 respectively, the root mean square error of calibration and validation is 0.171 and 0.220, and the correct recognition ratio for four brands vinegar is above 93.3%. The results of this study suggest that the combination of visible spectroscopy techniques and OSC might offer the possibility to classify the brands of vinegar without the need for costly and laborious analysis.