Abstract:Near infrared spectroscopy (NIRS) combined with partial least squares (PLS) method was used to achieve rapid quantitative analysis of crude fat and corase fiber in feed. The norris-williams derivation (NW) and multiplicative scatter correction (MSC) methods were used to pretreat the spectrum, and the monte carlo based uninformative variable elimination (MCUVE), variables combination population analysis (VCPA) and interval variable iterative space shrinkage approach (iVISSA) were used to select and optimize the variables of the spectrum. PLS was used for the establishment of the spectral calibration model, and the parameters of calibration set correlation coefficient (Rc), root mean square error of cross validation (RMSECV), prediction set correlation coefficient (Rp) and root mean square error of prediction (RMSEP) were used to evaluate the models. Compared with other pretreatment methods, the RMSECV and RMSEP values of the spectral model after MSC treatment decreased, while the Rc and Rp values increased. In the crude fat quantitative analysis model, the RMSECV and Rc of the original spectral model were 0.21 and 0.87, RMSEP and Rp were 0.20 and 0.88, and the number of variables (Vn) was 1501. After selecting variables by MCUVE method, RMSECV and Rc were 0.17 and 0.92, RMSEP and Rp were 0.19 and 0.89, and Vn was 400. For VCPA-PLS model, the RMSECV and Rc were 0.206 and 0.87, RMSEP and Rp were 0.25 and 0.81, and Vn was 12. For iVISSA-PLS model, the RMSECV and Rc were 0.21 and 0.86, RMSEP and Rp were 0.20 and 0.87, and Vn was 20. In the corase fiber model, the RMSECV and Rc of the original PLS model were 0.28 and 0.91, RMSEP and Rp were 0.23 and 0.95, and Vn was 1 501. After MCUVE selection, the RMSECV and Rc of the model were 0.23 and 0.95, RMSEP and Rp were 0.25 and 0.94, and Vn was 740. After VCPA selection, the RMSECV and Rc of the model were 0.27 and 0.91, RMSEP and Rp were 0.30 and 0.91, and Vn was 11. After iVISSA selection, the RMSECV and Rc of the model were 0.29 and 0.90, RMSEP and Rp were 0.27 and 0.93, and Vn was 20. The results showed that the MSC method could effectively improve the spectral quality and eliminate the spectral translation error; the MCUVE variable selection method could simplify the model to improve the model accuracy and stability, and establish the optimal model. In the crude fat quantitative analysis model, after the MSC-processed spectrum was selected by MCUVE, the remaining 400 were used to establish the PLS model, Rc and Rp were improved by 0.05 and 0.01 compared to the full-spectrum model, and the RMSECV and RMSEP were reduced to 0.17 and 0.19; The model selected by VCPA and iVISSA had almost the same result as the full-spectrum model, and its greatest feature was that only 12 and 20 variables were selected. In the corase fiber model, 740 variables selected by MCUVE were used to establish the PLS model with Rc and Rp of 0.95 and 0.94, RMSECV and RMSEP of 0.23 and 0.23, respectively; VCPA and iVISSA used 11 and 12 variables to establish the regression model, but its model results were all worse than the MCUVE model. The establishment of MSC-MCUVE-PLS quantitative analysis model using feed near-infrared spectroscopy could effectively quantify crude fat and corase fiber in feed.
Key words:Near infrared spectroscopy (NIRS); Feed; Monte carlo based uninformative variable elimination (MCUVE); Variables combination population analysis (VCPA); Interval variable iterative space shrinkage approach (iVISSA)
郝 勇,吴文辉,商庆园. 饲料中粗脂肪和粗纤维含量的近红外光谱快速分析[J]. 光谱学与光谱分析, 2020, 40(01): 215-220.
HAO Yong, WU Wen-hui, SHANG Qing-yuan. The Research on Quantitative Analysis of Feed Crude Fat and Corase Fiber Based on Near Infrared Spectroscopy and Variables Selection Methods. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2020, 40(01): 215-220.
[1] Sagrario M, Ana S, Adela MF, et al. Talanta, 2017, 162: 597.
[2] Patrica A H, Sarah N, Harry B C, et al. Journal of Equine Veterinary Science, 2018, 71: 13.
[3] Luisa M, Ilaria F, Giuseppina A, et al. Food Chemistry, 2018, 267: 240.
[4] Kelton S S, Anderson S S, Telma W L, et al. Journal of Computer Science, 2015, 11(4): 621.
[5] Hideyuki S, Junji M. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy, 2018, 192: 236.
[6] Asmund R, Frans B, Soren B E. Trebds in Analytical Chemistry, 2009, 28(10): 1201.
[7] Li C, Zhao T L, Li C, et al. Food Chemistry, 2017, 221: 990.
[8] Liu X W, Cui X Y, Yu X M, et al. Chinese Chemical Letters, 2017, 28: 1447.
[9] WU Jing-zhu, WANG Feng-zhu, WANG Li-li, et al(吴静珠, 汪凤珠, 王丽丽, 等). Spectroscopy and Spectral Analysis(光谱学与光谱分析), 2015, 35(1): 99.
[10] Yun Y H, Wang W T, Deng B C, et al. Analytica Chimica Acta, 2015, 862: 14.
[11] Yang Y N, Xie F F, Yan B, et al. Chemometrics and Intelligent Laboratory Systems, 2017, 170: 102.
[12] ZHAO Huan, HUAN Ke-wei, ZHENG Feng, et al(赵 环, 宦克为, 郑 峰, 等). Journal of Changchun University of Science and Technology(长春理工大学学报·自然科学版), 2016, 39(5): 51.
[13] Deng B C, Yun Y H, Liang Y Z, et al. Analyst, 2014, 139: 4836.
[14] Deng B C, Yun Y H, Ma P, et al. Analyst, 2015, 140: 1876.
[15] Ripoll G, Lobon S, Joy M. Meat Science, 2018, 143: 24.
[16] Fien D L, Elisabeth P, Hasna D, et al. Journal of Pharmaceutical and Biomedical Analysis, 2018, 151: 274.
[17] Pedro S S, Andreia S, Ana C, et al. Food Chemistry, 2018, 242: 196.