化学计量学法处理可见光谱定量测定纸浆卡伯值

本文将一元线性回归、多元线性回归和主成分回归分析的化学计量学方法应用于纸浆卡伯值的在线测量中,通过测量蒸煮液的吸光度来预测纸浆的卡伯值。结果显示：三种方法所确立的测量模型均具有较好的学习精度和预测精度,其中尤以多元线性回归模型同时具有较高的学习精度和预测鲁棒性。     

计学计量学法处理可见光谱定量测定纸浆卡伯值

本文将一元线性回归、多元线性回归和主成分回归分析的化学计量学方法应用于纸浆卡伯值的在线测量中,通过测量蒸煮 液的吸光度来预测纸浆的卡伯值。     

基于PCA降维模型的造纸废水处理过程软测量建模

针对造纸废水处理过程的复杂特性,本课题将主成分分析（PCA）与人工神经网络（ANN）和支持向量回归（SVR）相结合,构建出两种新的软测量模型：主成分分析-人工神经网络（PCA-ANN）和主成分分析-支持向量回归（PCA-SVR）。本课题将这两种软测量模型应用于造纸废水处理过程中出水化学需氧量（COD）和出水悬浮固形物（SS）浓度的预测。计算结果表明,PCA-ANN和PCA-SVR的预测效果均优于偏最小二乘、支持向量回归和人工神经网络3种常规软测量模型,并且PCA-ANN的预测效果最优。对于出水COD浓度预测,PCA-ANN的决定系数（R2）为0.984,均方误差（MSE）为1.892,较ANN分别优化了9.7%和71.5%。对于出水SS浓度预测,PCA-ANN的R2为0.762, MSE为0.228,较ANN分别优化了31.2%和58.7%。     

基于多元统计分析绝缘纸工频击穿强度影响因素的研究

为提高特高压变压器用绝缘纸的制造技术和质量水平、研究“材料-结构-性能”之间的相关性,提出一种基于灰色关联分析的绝缘纸工频击穿强度影响因素量化分析模型,并采用主成分分析和最佳子集选择方法进行多参数优化,构建绝缘纸工频击穿强度的多元线性回归模型。结果表明,绝缘纸工频击穿强度影响因素的灰色关联顺序为：纤维长度>细小纤维含量>纤维宽度>紧度>透气度>孔隙率>厚度。主成分分析表明,提取的3个主成分能够解释原来参数95.76%的信息,最佳子集选择将参数优化为3个参数;绝缘纸工频击穿强度的多元线性回归模型拟合度较高,模拟样本和验证样本预测结果的相对偏差基本在10%以内,表明回归模型具有良好的预测能力。     

电子鼻结合化学计量法对羊奶中蛋白质掺假的识别

利用电子鼻结合化学计量法对羊奶中的蛋白质掺假进行定性和定量的研究。用电子鼻检测掺入了不同蛋白质物质的羊奶,采用主成分分析、线性判别分析对电子鼻响应值进行定性分析,采用线性回归分析、Fisher判别分析以及K-最邻近值分析对电子鼻响应值进行定量分析。结果表明:主成分分析和线性判别分析都能够区分不同类别的掺假样品。线性回归分析的决定系数为84.5%,表明回归方程估测可靠程度较高。Fisher判别分析的原始分类的正确率达到100.0%,交叉验证的正确率为98.2%,说明其预测结果较好。K-最邻近值分析对训练集的分类正确率达到95.1%,对验证集的分类正确率为97.1%,说明模型的预测结果良好。说明应用电子鼻技术检测羊奶中的蛋白质掺假具有一定的可行性。     

基于近红外光谱技术检测红曲米中的红曲色素

采用近红外光谱技术结合化学计量学方法构建红曲米中红曲橙色素、红曲红色素、红曲黄色素的预测模型。分别采用多元线性回归(SMLR)、偏最小二乘回归(PLS)、主成分回归(PCR)构建所有色素组分的数学模型,以相关系数(R)、校正均方根误差(RMSEC)、预测均方根误差(RMSEP)、预测相对分析偏差(RPD)值来评价模型的综合性能。结果显示,MSC、SNV方法能够消除红曲米粉颗粒不均对光谱的散射影响;导数处理消除了基线漂移;对于红曲橙色素、红曲黄色素、红曲红色素三种模型均具有良好的稳定性;利用三种模型对未知红曲样品预测时,预测结果具有较高的线性,预测性能较好(RPD=2.86~5.39),可用于准确定量预测。结果表明近红外光谱技术可用于红曲色素的快速无损测定,为红曲米质量的智能化控制提供了新的途径。     

大豆油脂过氧化值的近红外光谱分析

首先采用碘量法测定油脂样品的过氧化值作为校正值,再使用近红外光谱仪采集油脂样品的光谱数据。在油脂过氧化值的定量分析中建立3 种校正模型,对比分析偏最小二乘回归(PLS)、多元线性回归(MLR)和主成分回归(PCR)的性能,确定最优模型为PLS。使用PLS 测定油脂过氧化值的性能结果为：校正集的相关系数0.916,预测集的相关系数0.922。     

响应面-主成分分析法优化马铃薯挂面工艺

本文以小麦粉、马铃薯全粉为主要原料,豆腐柴叶为辅料,采用单因素和Box-Behnken实验设计,响应面和主成分分析法优化马铃薯挂面加工工艺。结果表明:第1~第3主成分累计贡献率达到86.10%,足以描述马铃薯挂面质构、感官、烹调损失率和干物质吸水率综合反应的挂面品质。以主成分分析得到的规范化综合评分为响应值建立的二次多项式回归模型回归效果显著,拟合度较好(p0.0001,R2=0.9699)。偏最小二乘法回归分析预测马铃薯挂面最佳综合评分工艺参数为:马铃薯全粉添加量31%,豆腐柴汁液添加量9%,醒发时间31 min,醒发温度25℃,理论综合评分值达到0.9194,该条件下马铃薯挂面规范化综合评分达0.9116,与模型预测值接近,表明以主成分分析得到的规范化综合评分为响应值建立的回归模型具有良好的预测能力。     

基于特征氨基酸指纹图谱的白燕窝识别掺假模型建立与验证

测定各类燕窝和掺假样品的氨基酸组成,通过主成分分析构建白燕窝鉴别的综合评价模型,运用多变量统计分析方法,初步建立燕窝掺假模型,并进行验证。结果表明,前两个主成分提取了原来18个指标87.75%的信息;利用综合评价模型计算出各类样品的主成分得分,燕窝样品在主成分得分图中分布的位置较集中,掺假燕窝样品则较分散;以掺假物质含量作为自变量,综合评价得分作为因变量,进行线性回归分析,单一物质掺假的线性模型和混合掺假线性模型的线性系数R20.98。模型验证误差平均值为3.25%,说明氨基酸指纹图谱指纹谱图鉴别法是一种有效且可靠性高的鉴别白燕窝的方法。     

显微高光谱技术检测肌细胞中超氧化物歧化酶活力的方法研究

利用可见显微高光谱技术对羊肉肌细胞中的超氧化物歧化酶(Superoxide dismutase,SOD)活力进行检测。通过显微高光谱系统(380~980 nm)采集223个显微图像,根据样本光谱的反射率提取感兴趣区域并结合SOD酶活力建立模型。对原始光谱结合偏最小二乘回归模型,进行样本集划分及多种光谱预处理的模型对比分析,优选出多元散射校正为预处理方法,采用6种方法提取特征波长,并根据特征波长建立偏最小二乘回归、多元线性回归、主成分回归三种模型。结果显示,建立基于竞争性自适应重加权法挑选特征波长的多元线性回归模型最优,预测集的相关系数和均方根误差分别为0.8351和21.3578 U/mg·prot。采用显微高光谱成像技术对肌细胞内超氧化物歧化酶活力的检测是具有可行性的。     

Using eigenvalues as variance priors in the prediction of genomic breeding values by principal component analysis

Genome-wide selection aims to predict genetic merit of individuals by estimating the effect of chromosome segments on phenotypes using dense single nucleotide polymorphism (SNP) marker maps. In the present paper, principal component analysis was used to reduce the number of predictors in the estimation of genomic breeding values for a simulated population. Principal component extraction was carried out either using all markers available or separately for each chromosome. Priors of predictor variance were based on their contribution to the total SNP correlation structure. The principal component approach yielded the same accuracy of predicted genomic breeding values obtained with the regression using SNP genotypes directly, with a reduction in the number of predictors of about 96% and computation time of 99%. Although these accuracies are lower than those currently achieved with Bayesian methods, at least for simulated data, the improved calculation speed together with the possibility of extracting principal components directly on individual chromosomes may represent an interesting option for predicting genomic breeding values in real data with a large number of SNP. The use of phenotypes as dependent variable instead of conventional breeding values resulted in more reliable estimates, thus supporting the current strategies adopted in research programs of genomic selection in livestock.     

基于LF-NMR弛豫特性的煎炸油总极性化合物含量定量建模方法

将低场核磁共振（low field nuclear magnetic resonance,LF-NMR）分析技术应用于煎炸油脂总极性化合物（total polar compounds,TPC）含量的预测。采用柱层析方法测定油脂样品的TPC含量作为测定值,采集油脂样品的LF-NMR弛豫特性（峰起始时间T21、T22、T23相应的峰面积比例S21、S22、S23、单组分弛豫时间T2W）,分别利用向后筛选多元回归分析、主成分回归分析和偏最小二乘回归分析建立LF-NMR弛豫特性与TPC含量的回归方程,比较3 种模型的校正集和预测集的决定系数与均方根误差,最终确定最优模型为偏最小二乘回归模型。应用此模型预测预测集样品TPC含量,决定系数R2可达0.928,预测集均方根误差为0.568%,模型稳定。     

Nondestructive measurement of soluble solid content of navel orange fruit by visible-NIR spectrometric technique with PLSR and PCA-BPNN

A relationship was established between the soluble solid content (SSC) of navel orange fruit determined by destructive measurement and visible-near infrared diffuse reflectance spectra in the wavelength range of 350-1800 nm. Multiplicative scatter correction (MSC) and standard normal variate correction (SNV) were applied to the spectra, partial least squares regression (PLSR) and back propagation neural network (BPNN) based on principal component analysis (PCA) were used to develop the models for predicting the SSC of intact navel orange fruit. Thirty-eight unknown samples were used to evaluate the performance of these models. The principal component analysis-back propagation (PCA-BPNN) method with MSC spectral pretreatment obtain the best predictive results, resulting in correlation coefficient, root mean square error of prediction (RMSEP), average difference between predicted and measured values (Bias) of 0.90, 0.68 °Brix and 0.16 °Brix, respectively. Experimental results indicate that PCA-BPNN is a suitable tool to model the non-linear complex system, with additional advantages over PLSR, and the vis/NIR spectrometric technique can be used for measuring the SSC of intact navel orange fruit, nondestructively.     

Prediction of genomic breeding values for dairy traits in Italian Brown and Simmental bulls using a principal component approach

The large number of markers available compared with phenotypes represents one of the main issues in genomic selection. In this work, principal component analysis was used to reduce the number of predictors for calculating genomic breeding values (GEBV). Bulls of 2 cattle breeds farmed in Italy (634 Brown and 469 Simmental) were genotyped with the 54K Illumina beadchip (Illumina Inc., San Diego, CA). After data editing, 37,254 and 40,179 single nucleotide polymorphisms (SNP) were retained for Brown and Simmental, respectively. Principal component analysis carried out on the SNP genotype matrix extracted 2,257 and 3,596 new variables in the 2 breeds, respectively. Bulls were sorted by birth year to create reference and prediction populations. The effect of principal components on deregressed proofs in reference animals was estimated with a BLUP model. Results were compared with those obtained by using SNP genotypes as predictors with either the BLUP or Bayes_A method. Traits considered were milk, fat, and protein yields, fat and protein percentages, and somatic cell score. The GEBV were obtained for prediction population by blending direct genomic prediction and pedigree indexes. No substantial differences were observed in squared correlations between GEBV and EBV in prediction animals between the 3 methods in the 2 breeds. The principal component analysis method allowed for a reduction of about 90% in the number of independent variables when predicting direct genomic values, with a substantial decrease in calculation time and without loss of accuracy.     

Predicting Protein Functionality with Artificial Neural Networks: Foaming and Emulsifying Properties

Using physicochemical properties of 11 food-related proteins, artificial neural networks (ANN) were developed for predicting foam capacity and stability and the emulsion activity index. The prediction accuracy of ANN was compared to that of principal component regression (PCR) models. ANN had better prediction ability than PCR, especially after cross-validation. For foam stability, PCR did not generate a significant model. ANN and PCR models indicated that fluorescence probe hydrophobicity (measured using cispsrinaric acid) and other properties, such as viscosity, surface tension and net charge were important in predicting foam capacity and stability.     

Genetic parameter estimates for body conformation traits using composite index,principal component,and factor analysis

Information about genetic parameters is population specific and it is crucial for designing animal breeding programs and predicting response to selection. This study was carried out to estimate the genetic parameters for 23 body conformation traits of 45,517 Chinese Holstein reared in Eastern China from 1995 to 2017 with the Bayesian inference method using a linear animal mixed model. The methods to integrate these traits included (1) using the composite index from the Dairy Association of China and (2) applying principal component analysis and factor analysis to explore the relationship between the conformation traits. Estimates of heritability using the composite index were low (0.04; feet and legs) to moderate (0.23; body capacity). Strong genetic correlations were observed between the individual body conformation traits. Both principal components (1 to 7; eigenvalues ≥ 1) and latent factors (1 to 7; eigenvalues ≥ 1) explained 60.37% of total variability. Principal component 1 and factor 1 accounted for the traits that are usually associated with milk production. Moderate to low heritability were estimated through multi-trait analysis for principal components (from 0.07 to 0.21) and latent factors (from 0.07 to 0.23). Genetic correlations among the 2 multivariate techniques are typically lower compared with the one existing among the measured traits. Results from these analyses suggest the possibility of using both principal component analysis and factor analysis in morphological evaluation, simplifying the information given by the body conformation traits into new variables that could be useful for the genetic improvement of the Chinese Holstein population. This information could also be used to avoid analyzing large number of correlated traits, thereby improving precision and reducing computation burdens to analyze large and complex data.     

PLS法和PCA法在近红外光谱定量分析中的应用研究

利用啤酒的近红外光谱数据比较了 PLS(偏最小二乘法,partial-Squares)和 PCA(主成分回归法,principal componentregression)两种方法在近红外光谱定量分析中的应用。并应用所建模型预测了 21 个啤酒样品麦芽的含量,结论为两种方法均适合近红外光谱定量分析,PLS 法所得预测结果准确度更高。     

Measurement of yogurt internal quality through using Vis/NIR spectroscopy

A nondestructive optical method for determining the sugar content and acidity of yogurt was investigated. Three types of preprocessing, S. Golay smoothing with multiplicative scatter correction (S. Golay smoothing with MSC), S. Golay 1st-Der and wavelet package transform (WPT), were used before the data were analyzed with chemometrics methods of partial least square (PLS). Spectral data sets as the logarithms of the reflectance reciprocal were analyzed to build a best model for predicting the sugar content and acidity of yogurt. A model using preprocessing of WPT with a correlation coefficient of 0.91 and 0.90, a root mean square error of prediction (RMSEP) of 0.36 and 0.04 showed an excellent prediction performance to sugar content and acidity. S. Golay smoothing with MSC was also finer, combined with the calibration and validation results. S. Golay 1st-Der was the worse preprocessing method in this experiment. In the paper, a multivariate calibration method of principal component artificial neural network (PC-ANN) was also established. In PC-ANN models, the scores of the principal components were chosen as the input nodes for the input layer of ANN. After adjusting the number of input nodes (principal components), hidden nodes, as well as learning rate and momentum of the network, a model with a correlation coefficient of 0.92 and 0.91, a root mean square error of prediction (RMSEP) of 0.33 and 0.04 showed an excellent prediction performance on sugar content and acidity. At the same time, the sensitive wavelengths corresponding to the sugar content and acidity of yogurt were proposed on the basis of regression coefficients by PLS.