该文以近红外光谱分析技术快速测定菠萝啤中果汁含量为目的,采用了后向间隔偏最小二乘(backward interval partial least squares,Bi-PLS)、组合间隔偏最小二乘(synergy interval partial least squares,Si-PLS)以及遗传算法(genetic algorithm,GA)提取特征波长以提高模型性能。研究结果表明,基于Si-PLS提取的特征波长结合偏最小二乘法(partial least squares, PLS)建立的定量分析模型效果最好,从原始光谱范围4 000~10 000 cm-1内筛选出3个特征光谱区间,分别为(4 484~4 960,5 600~6 051,7 844~8 080) cm-1,共94个特征变量,比原始1 501个波长变量减少了93.7%,验证集的均方根误差和决定系数分别为0.18%、0.89,范围误差比为3.17。实验结果表明,近红外光谱分析技术用于测定果味啤中的果汁含量是可行的,这为快速高效测定菠萝啤果汁含量提供了一种方法依据。
Near-infrared spectroscopy technique was applied to quickly determine the juice content in pineapple beer. Backward interval partial least squares (Bi-PLS), synergy interval partial least squares (Si-PLS) and genetic algorithm (GA) were used to extract characteristic wavelengths prior to build PLS regression models. The performance of PLS model was evaluated by the Decision coefficient (Rp2), the root mean square error (RMSEP) and the ratio of performance to standard deviate (RPD) in the prediction set. Among the methods used, the Si-PLS was found to be superior to other methods. The predicted root mean square error (RMSEP), determination coefficients for prediction sets (RP2) and ratio of performance to standard deviate (RPD) was 0.18%, 0.89 and 3.17 respectively. The characteristic spectral intervals are 4 484-4 960, 5 600-6 051 and 7 844-8 080 cm-1. And a total of 94 characteristic variables which decreaded by 93.7% than the previous 1 501 wavelength variable. The experimental results indicated that it is feasible to measure the juice content of pineapple beer by near-infrared spectroscopy. This study provided a method for the rapid and efficient determination of pineapple juice content.
[1] 杜冰, 姚汝华. 果汁啤酒的生产与发展趋势[J]. 现代食品科技, 2001, 17(1): 21-22.
[2] 尚静, 张艳. 猕猴桃菠萝啤的研制与品质评价[J]. 农产品加工, 2018, 467(11): 14-16.
[3] 汪中世. 蓝莓菠萝啤研制及其稳定性研究[D]. 合肥:安徽农业大学, 2015.
[4] 宋淑红, 徐慧琴. 6°P菠萝菠萝啤的生产[J]. 农产品加工(学刊), 2014(13): 35-36.
[5] DONG D, LU C, CAO Y, et al. Present situation and development strategy of fruit beer in China[J]. Journal of Food Safety & Quality, 2015,6(7):2 626-2 632.
[6] 董德武, 卢存龙, 曹圆圆, 等. 我国果汁啤酒生产现状及发展战略[J]. 食品安全质量检测学报, 2015(7):2 626-2 632.
[7] 张亮, 赵长新. 菠萝啤与果汁组分差异的比较[J]. 食品与发酵工业, 2006, 32(9): 149-152.
[8] 谷如祥, 赵武奇, 石珂心,等. 近红外光谱测定苹果饮料中原果汁含量[J]. 食品工业科技, 2013, 34(20): 75-77.
[9] 牛灿杰, 张慧, 王爽,等. 浓缩还原橙汁饮料中果汁含量快速检测方法探究[J]. 食品工业科技, 2015, 36(8): 72- 77.
[10] 褚小立. 近红外光谱分析技术实用手册[M]. 机械工业出版社, 2016.
[11] CHU Xiaoli,LU Wanzhen. Research and application progress of near infrared spectroscopy analytical technology in China in the past five years[J]. Guang Pu Xue Yu Guang Pu Fen XI, 2014, 34(10): 2 595-2 605.
[12] 张世芝, 胡树青, 张明锦. 基于回归系数的变量筛选方法用于近红外光谱分析[J]. 计算机与应用化学, 2012, 29(2): 227-230.
[13] NADLER B, COIFMAN R R. The prediction error in CLS and PLS: The importance of feature selection prior to multivariate calibration [J]. Journal of Chemometrics, 2010, 19(2): 107- 118.
[14] RADY A, GUYER D. Utilization of visible/near-infrared spectroscopic and wavelength selection methods in sugar prediction and potatoes classification[J]. Journal of Food Measurement & Characterization, 2015, 9(1): 20- 34.
[15] MOHAMMADREZA K, MOBEDI H, MOBEDI E, et al. Quantitative determination of naltrexone by attenuated total reflectance – FTIR spectrometry using partial least squares (PLS) wavelength selection [J]. Spectroscopy, 2015, 23(2): 113-121.
[16] SPIEGELMAN C H, MCSHANE M J, GOETZ M J, et al. Theoretical justification of wavelength selection in PLS calibration: Development of a new algorithm [J]. Analytical Chemistry, 1998, 70(1): 35- 44.
[17] CHU Xiaoli, XU Yupeng, LU Wanzhen. Research and application progress of chemometrics methods in near infrared spectroscopic analysis[J]. Chinese Journal of Analytical Chemistry, 2008, 36(5): 702-709.
[18] 张严, 谢岩黎, 孙淑敏. 近红外光谱结合化学计量学方法在油脂检测中的应用[J]. 粮食与油脂, 2015, 225(1): 66-68.
[19] 樊双喜, 钟其顶, 李国辉,等. 近红外光谱法快速检测黄酒的酒精度、总糖和总酸[J]. 中国酿造, 2015, 34(2): 135-138.
[20] 张亮, 赵长新. 果啤与果汁组分差异的比较[J]. 食品与发酵工业, 2006, 32(9): 149-152.
[21] 刘桂松. Vis-NIR光谱判别分析的几类模式识别方法研究[D]. 广州:暨南大学, 2015.
[22] QU FANGFANG, REN DONG, HOU JINJIAN, et al. The characteristic spectral selection method based on forward and backward interval partial least squares[J]. Spectroscopy and Spectral Analysis, 2016, 36(2): 593-598.
[23] SHARIATI-RAD M, MASOUMEH H. Selection of individual variables versus intervals of variables in PLSR [J]. Journal of Chemometrics, 2010, 24(2): 45-56.
[24] DENG BAICHUAN, YUN YONGHUAN, MA PAN, et al. A new method for wavelength interval selection that intelligently optimizes the locations, widths and combinations of the intervals [J]. The Analyst, 2015, 140(6): 1 876-1 885.
[25] AHLINDER J, NORDGAARDA. Chemometrics comes to court: Evidence evaluation of chem-bio threat agent attacks[J]. Journal of Chemometrics, 2015, 29(5): 267-276.
[26] 雷英杰, 张善文. MATLAB遗传算法工具箱及应用[M]. 西安:西安电子科技大学出版社, 2014.
[27] MARIANI, NATHLIA C T, et al. Predicting soluble solid content in intact jaboticaba [Myrciaria jaboticaba (Vell.) O. Berg] fruit using near-infrared spectroscopy and chemometrics [J]. Food Chemistry, 2014, 15(9): 458-462.
[28] 马永杰, 云文霞. 遗传算法研究进展[J]. 计算机应用研究, 2012, 29(4): 1 201-1 206.
[29] LU I R R, KWAN E, THOMAS D R, et al. Two new methods for estimating structural equation models: An illustration and a comparison with two established methods. [J]. International Journal of Research in Marketing, 2011, 28(3): 258-268.
[30] LEARDIR. Application of genetic algorithm-PLS for feature selection in spectral data sets [J]. Journal of Chemometrics, 2010, 14(5-6): 643-655.
[31] HUANG GUANGZAO, RUAN XIUKAI, CHEN XIAOJING, et al. A segmented PLS method based on genetic algorithm [J]. Analytical Methods, 2014, 6(9): 2 900-2 908.
