PLS-BP法近红外光谱技术同时测定鲜乳中四种主成分

基于近红外光谱技术,将偏最小二乘法(Partial Least Squares,PLS)和单隐层的反向传播网络(Back-Propagation Network,BP)联用并测定了鲜乳中4种主成分含量.用PLS法将原始数据压缩为主成分,取前3个主成分的14个数据输入网络,以Kolmogorov定理为依据,经过实验确定中间层的神经元个数为29,初始训练迭代次数为1000,建立了脂肪、蛋白质、乳糖、牛乳总固体4种主成分含量的预测校正模型.PLS-BP模型对样品4个组分含量的预测决定系数(R2)分别为:0.961、0.974、0.951、0.997;本研究为近红外光谱技术在鲜乳多组分快速检测提供了新思路.     

偏最小二乘法测定复方乙酰水杨酸片中的有效成分

将偏最小二乘法(PLS)与近红外漫反射光谱法相结合,对复方乙酰水杨酸片进行无损非破坏定量分析．建立了最佳的数学校正模型,比较了样品中3种有效成分(乙酰水杨酸、非那西丁和咖啡因)同时测定和单独测定时的主成分数对PLS定量预测能力的影响,预测了未知样品。3种有效成分同时测定和单独测定建立的PLS模型具有相同的主成分数,PLS预报浓度与参考浓度具有相近的标准偏差,说明用PLS法同时测定3种组分的含量是可行的。     

近红外光谱结合CARS变量筛选方法用于液态奶中蛋白质与脂肪含量的测定

采用CARS(Competitive adaptive reweighted sampling)变量筛选方法建模,显著提高了液态奶中蛋白质与脂肪近红外模型的预测精度。用蒙特卡罗采样(Monte-Carlo sampling)方法先剔除奇异样本,再对光谱进行中心化与Karl Norris滤波降噪处理,通过CARS方法筛选出与样本性质密切相关的变量,建立预测蛋白质与脂肪含量的偏最小二乘法(PLS)校正模型,并与未选变量的PLS模型进行比较。以定标集相关系数(r2)及交互验证均方残差(RMSECV)和预测误差均方根(RMSEP)作为判定依据,确定了蛋白质与脂肪的最佳建模条件。蛋白质与脂肪校正模型的相关系数分别为0.975 0、0.995 1,RMSECV分别为0.194 8、0.136 3,RMSEP分别为0.113 3、0.140 1,预测结果优于未选变量的PLS模型及其他选变量方法,有效简化了模型,适于液态奶中脂肪和蛋白质的快速、无损检测。     

基于PLS-自组织竞争神经网络近红外光谱技术对鲜乳和掺假乳的检测方法研究

通过偏最小二乘法(partial least squares,PLS)与人工神经网络(artificial neural networks,ANN)联用对鲜乳和掺有植物奶油的牛乳建立识别模型.用PLS法对原始数据进行主成分压缩,采用自组织竞争神经网络建模.取前3个主成分的21个吸收峰值输入网络,学习参数为0.05,网络训练迭代次数为200,模型鉴别准确率达100%.其次建立了植物奶油掺假量的定量检测PLS模型,并采用交互校验和外部检验考察模型的可靠性,模型的校正相关系数为0.996 3,均方估计残差(RMSEC)为0.110;交互校验均方残差(RMSECV)为0.142;应用所建PLS模型对样品中植物奶油添加量进行预测,并对预测值与真值进行配对t检验,结果表明两者差异均不显著.     

基于灰狼算法的近红外光谱变量选择方法研究

基于群体智能的灰狼优化(GWO)算法具有参数少、结构简单、易于实现的优点,但在光谱领域的应用较少。该研究将GWO算法引入近红外光谱的变量筛选中,以玉米数据为例,考察了GWO算法中狼群性能、迭代次数、狼群数量及运算效率,并建立了偏最小二乘(PLS)模型对玉米样品中蛋白质、脂肪、水分以及淀粉含量的测定。结果显示,GWO算法运算效率很高,经过参数调优后建立PLS模型,其蛋白质、脂肪、水分及淀粉的保留变量数分别为19、19、14、34,预测均方根误差(RMSEP)从全波长PLS建模的0.245 8、0.122 4、0.339 8、1.105 8分别下降到0.147 7、0.080 1、0.176 2、0.739 8,分别下降了40%、35%、48%、33%,相关系数也相应地提高。因此,GWO算法不仅优化速度快,选择变量数少,还可以显著提高PLS模型的预测精度,是一种近红外光谱变量选择的有效方法。     

支持向量机回归法在近红外光谱测定植物纤维原料中甲氧基含量中的应用

以26个植物纤维原料为实验材料,由20个样品作校正样品,采用径向基核函数方法对纤维原料中甲氧基含量与纤维原料样品近红外光谱进行支持向量机(SVM)回归建模.以所建SVM回归模型对6个纤维原料样品中甲氧基含量进行预测,回归模型的预测结果与采用改良的维伯克法确定的甲氧基含量的相关系数为0.977,预测样本集的标准偏差为0.43.将SVM回归模型的预测效果与PLS回归模型的预测结果进行比较,所建近红外光谱测定植物纤维原料中甲氧基含量的SVM回归模型可用于实际植物纤维原料样品的定量分析,且具有较好的分析效果.     

近红外光谱主成分-所有可能回归法定量分析烤烟、小麦样品中的组分含量

利用主成分-所有可能回归法,建立了烤烟、小麦样品不同组份的近红外光谱定量分析模型。结果表明,烤烟样品的总糖、还原糖以及小麦样品的蛋白质含量的预测模型均有好的定量分析结果,且其预测结果与PLS法预测结果相当。     

蛋白质二级结构预测的人工神经网络方法研究

本文比较了五种神经网络方法预测蛋白质二级结构的准确率,并做出初步评价。五种神经网络分别是:误差反传前向网络(BP),径向基函数网络(RBF),广义回归神经网络(GRNN),串并联叠层网络(CF),Elman网络(ELM)。结果显示:GRNN的预测准确率达85.7%,优于其它网络。本文还讨论了训练集样本数及参数的优化对GRNN预测准确率的影响。     

基于最小二乘支持向量机的油页岩含油率近红外光谱分析

为了提高油页岩含油率近红外光谱分析建模的预测精度和稳定性,开展了基于最小二乘支持向量机(LS-SVM)建模方法的对比研究.采用主成分-马氏距离(PCA-MD)和基于蒙特卡洛采样(MCS)2种方法进行了奇异样本的检测,采用径向基核函数的LS-SVM、偏最小二乘(PLS)和反向传播神经网络(BPANN)3种方法进行建模方法对比.结果表明,对于64个油页岩岩芯样本,与PCA-MD方法相比,采用MCS方法剔除奇异样本后所建PLS模型的预测精度提高了28%.对于MCS方法剔除奇异样本后的58个样品,采用KennardStone法划分了44个样品的校正集和14个样品的预测集,采用2阶导数和标准化预处理方法,建立了100个LS-SVM的校正模型,模型的预测决定系数R2平均值达到0.90以上,高于PLS和BPANN模型的对应值;且R2的变化量(0.02)小于BPANN模型的对应值(0.32).因此,MCS奇异样本检测结合LS-SVM方法可提高油页岩含油率样本建模的精度和稳定性.     

复杂样品近红外光谱定量分析模型的构建方法

针对复杂样品近红外光谱分析中校正集的设计问题, 探讨了标准样品参与复杂样品建模的可行性. 通过标准样品和复杂基质样品共同构建的偏最小二乘(PLS)模型, 考察了波段筛选和建模参数对预测结果的影响. 结果表明, 采用PLS方法建立定量模型时, 校正集样品性质应该尽量与预测集样品相似, 当样品的性质相差较大时, 适当增加校正集样品的差异性可使模型具有更强的预测能力. 同时, 波段优选对提高预测结果的准确性具有重要的意义.     

A Partial Least Squares‐Based Consensus Regression Method for the Analysis of Near‐Infrared Complex Spectral Data of Plant Samples

Abstract

A consensus regression approach based on partial least square (PLS) regression, named as cPLS, for calibrating the NIR data was investigated. In this approach, multiple independent PLS models were developed and integrated into a single consensus model. The utility and merits of the cPLS method were demonstrated by comparing its results with those from a regular PLS method in predicting moisture, oil, protein, and starch contents of corn samples using the NIR spectral data. It was found that cPLS was superior to regular PLS with respect to prediction accuracy and robustness.     

Near-infrared reflectance spectroscopy and multivariate calibration techniques applied to modelling the crude protein, fibre and fat content in rapeseed meal

Near-infrared reflectance spectroscopy (NIRS) is often applied when a rapid quantification of major components in feed is required. This technique is preferred over the other analytical techniques due to the relatively few requirements concerning sample preparations, high efficiency and low costs of the analysis. In this study, NIRS was used to control the content of crude protein, fat and fibre in extracted rapeseed meal which was produced in the local industrial crushing plant. For modelling the NIR data, the partial least squares approach (PLS) was used. The satisfactory prediction errors were equal to 1.12, 0.13 and 0.45 (expressed in percentages referring to dry mass) for crude protein, fat and fibre content, respectively. To point out the key spectral regions which are important for modelling, uninformative variable elimination PLS, PLS with jackknife-based variable elimination, PLS with bootstrap-based variable elimination and the orthogonal partial least squares approach were compared for the data studied. They enabled an easier interpretation of the calibration models in terms of absorption bands and led to similar predictions for test samples compared to the initial models.     

Averaged and weighted average partial least squares

Two alternative partial least squares (PLS) methods, averaged PLS and weighted average PLS, are proposed and compared with the classical PLS in terms of root mean square error of prediction (RMSEP) for three real data sets. These methods compute the (weighted) average of PLS models with different complexity. The prediction abilities of the alternative methods are comparable to that of the classical PLS but they do not require to determine how many components should be included in the model. They are also more robust in the sense that the quality of prediction depends less on a good choice of the number of components to be included. In addition, weighted average PLS is also compared with the weighted average part of LOCAL, a published method that also applies weighted average PLS, with however an entirely different weighting scheme.     

Orthogonal projection to latent structures solution properties for chemometrics and systems biology data

Partial least squares (PLS) is a widely used algorithm in the field of chemometrics. In calibration studies, a PLS variant called orthogonal projection to latent structures (O‐PLS) has been shown to successfully reduce the number of model components while maintaining good prediction accuracy, although no theoretical analysis exists demonstrating its applicability in this context. Using a discrete formulation of the linear mixture model known as Beer's law, we explicitly analyze O‐PLS solution properties for calibration data. We find that, in the absence of noise and for large n, O‐PLS solutions are simpler but just as accurate as PLS solutions for systems in which analyte and background concentrations are uncorrelated. However, the same is not true for the most general chemometric data in which correlations between the analyte and background concentrations are nonzero and pure profiles overlap. On the contrary, forcing the removal of orthogonal components may actually degrade interpretability of the model. This situation can also arise when the data are noisy and n is small, because O‐PLS may identify and model the noise as orthogonal when it is statistically uncorrelated with the analytes. For the types of data arising from systems biology studies, in which the number of response variables may be much greater than the number of observations, we show that O‐PLS is unlikely to discover orthogonal variation whether or not it exists. In this case, O‐PLS and PLS solutions are the same. Copyright © 2011 John Wiley & Sons, Ltd.     

应用遗传算法和PLS的近红外光谱预测玉米中淀粉含量的研究

以普通玉米籽粒为试验材料,在应用遗传算法结合偏最小二乘回归法对近红外光谱数据进行特征波长选择的基础上,应用偏最小二乘回归法建立了特征波长测定玉米籽粒中淀粉含量的校正模型．试验结果表明,基于11个特征波长所建立的校正模型,其校正误差（RMSEC）、交叉检验误差（RMSECV）和预测误差（RMSEP）分别为0.30％、0.35％和0.27％,校正数据集和独立的检验数据集的预测值与实际测定值之间的相关系数分别达到0.9279和0.9390,与全光谱数据所建立的预测模型相比,在预测精度上均有所改善,表明应用遗传算法和PLS进行光谱特征选择,能获得更简单和更好的模型,为玉米籽粒中淀粉含量的近红外测定和红外光谱数据的处理提供了新的方法与途径．     

偏最小二乘-反向传播-近红外光谱法同时测定饲料中4种氨基酸

偏最小二乘与人工神经网络联用对70个饲料样品建立起天门冬氨酸(Asp)、谷氨酸(Glu)、丝氨酸(Ser)和组氨酸(His)4种氨基酸含量的预测校正模型,以样品平行扫描光谱验证校正模型预测的准确性和重现性。用偏最小二乘法将原始数据压缩为主成分,采用单隐层的反向传播网络建模。取前3个主成分的12个数据输入网络,以Kolmogorov定理为依据,经过实验确定中间层的神经元个数为25,初始训练迭代次数为1000。偏最小二乘-反向传播网络模型对样品4个组分含量的预测决定系数(R2)分别为:0.981、0.997、0.979、0.946;样品平行扫描光谱预测值的标准偏差分别为:0.020、0.029、0.017、0.023。本研究为近红外快速检测在组分含量较低的样品实现多组分同时测定提供了思路。     

Components Analysis of Biochar Based on Near Infrared Spectroscopy Technology

This study aims to establish a rapid quantitative analysis method for biochar based on near infrared spectroscopy (NIRS) technology. Near infrared spectra of 163 samples in the 10000–3800 cm^–1 (1000–2632 nm) range were collected, and the contents of fixed carbon (FC), volatile matter (VM) and ash of samples were also analyzed. A partial least square (PLS) model for FC, VM and Ash was established after the model spectral ranges were optimized, the optimal factors were determined, and the raw spectra were pretreated by multiple scatter correction and second derivative (MSC + SD) method. Finally, the prediction performance of predictive model was evaluated. The results showed that the PLS model had a good prediction ability, and the predicted coefficient R²_p of actual values vs prediction values for FC, VM and ash were 0.9423, 0.9517 and 0.9265, respectively. Root mean square error of prediction (RMSEP) was 0.1074, 0.1201 and 0.1243, and ratios of prediction to deviation (RPD) were 3.51, 4.28 and 2.03, respectively. The PLS model had good accuracy and precision for both of FC and VM, and could be used as a quantitative method for FC and VM contents analysis. Nevertheless, PLS model need to improve the precision for Ash analysis according to RPD value. This method provides a fast and effective technical means for the quantitative analysis of biochar components.