基于最小角回归结合一元线性直接校正法的近红外光谱模型传递方法

针对近红外光谱分析技术中模型通用性较差的问题,提出了一种新的模型传递方法——最小角回归结合一元线性直接校正法(Least angle regression combined simple linear regression direct standardization,LARSLRDS)。该方法首先采用小波变换对样品光谱数据进行预处理,然后利用LAR实现样品全谱区光谱特征波长点的筛选,最后利用SLRDS对筛选出来的变量进行校正。采用汽油和药品样本的近红外光谱数据验证LAR-SLRDS性能,汽油数据集C7、C8、C9和C10成分的光谱差异为0. 002 8、0. 002 7、0. 002 6和0. 002 7,预测标准差为0. 410 6、0. 849 2、1. 034 9和1. 215 8;药品数据集活性、硬度和重量成分的光谱差异为0. 030 0、0. 031 8和0. 033 6,预测标准差为1. 933 8、0. 440 2和2. 130 9。结果表明,LAR-SLRDS算法不仅能够消除主、从仪器光谱之间存在的差异,实现模型传递,而且能够提高PLS定量模型的准确性和稳定性,具有广泛的应用潜力。     

小波变换-分段直接校正法用于近红外光谱模型传递研究

提出了一种新的传递算法(WT-PDS)———小波变换-分段直接校正法,并详细讨论了模型传递参数和传递结果。首先利用小波变换对光谱进行压缩处理,采用PDS算法消除不同仪器之间压缩数据的差异,最后利用经校正的压缩数据进行分析,实现模型传递。本方法能够扣除不同仪器之间的大部分差异,大幅度改善分析精度。传递后模型分析精度与源机模型稳健性紧密相关。如果源机模型稳健性强,则能够实现不同仪器之间的共享。本方法能够实现源机的0#轻柴十六烷值、凝点、馏出温度;-10#轻柴十六烷值、凝点以及-10#军柴凝点和馏出温度共10个模型在5台仪器之间共享,简化了建模的成本。与传统的PDS相比,WT-PDS方法具有传递和建模变量少、速度快、光谱校正性能高等优点,而其模型分析精度与传统PDS基本一致。     

基于稳定一致波长筛选的无标样近红外光谱模型传递方法

测量环境及仪器间光谱信号的差异导致近红外光谱模型从主机传递到从机后,经常会产生过大误差。本研究提出了一种基于稳定一致波长筛选的无标样近红外模型传递方法(Screening stable and consistent wavelengths,SSCW),剔除主从仪器间差谱的标准偏差大于样品精密度测试光谱标准偏差的波长,以及精密度测试偏差过大的波长,筛选出仪器间光谱信号一致性好且稳定的波长建立近红外光谱定标模型。分别以玉米和黄芩样本集对本算法的有效性进行了检验。结果表明,SSCW模型传递后对从机样品的预测均方根残差RMSEP较全波长PLS模型直接传递结果小一个量级,大部分情况下优于分段直接校正算法(Piecewise direct standardization,PDS)的结果和文献报道的无标样模型传递结果。本方法具有传递性能好、模型参数少、稳健等优点,在不同仪器间可实现近红外光谱模型的无标样传递。     

二进制蜻蜓算法-直接校正算法优选标样集在小麦粉蛋白质近红外模型传递中的应用

为了实现小麦粉蛋白质含量近红外分析模型的传递，探究二进制蜻蜓算法(Binary Dragonfly Algorithm, BDA)与直接校正算法(Direct Standardization, DS)相结合构成的BDA-DS算法挑选标样集对模型传递结果的影响。以棱光S450光栅型近红外光谱仪为主机，NeoSpectra Micro傅里叶变换型近红外光谱仪为从机，采集了126个小麦粉的近红外光谱，用偏最小二乘回归法建立了主机近红外光谱与小麦粉蛋白质的关联模型。经BDA-DS算法模型传递后，主机模型对从机样品预测决定系数为0.9812,预测标准偏差为0.1838,从机与主机的光谱集合平均马氏距离由22.34下降到1.40,均接近于主机模型精度水平。该研究同时与采用Kennard/Stone(K/S)挑选标样集再结合DS构成的传统K/S-DS算法进行了对比，结果表明：相对于K/S-DS算法，BDA-DS算法挑选出较少的标样集就能表征仪器的差异，有效地提高了主机模型对从机样品的预测精度，为近红外模型传递提供了一种更加有效的标样集选择方法。     

基于光谱比值分析的无标样近红外模型传递方法

测量环境及光谱仪台间差异导致近红外光谱(NIRS)模型传递到从机后,常产生较大误差。该文使用标准正态变量变换(SNV)+微分处理光谱消除光谱散射和基线漂移的影响,提出通过仪器间光谱信号比值分析筛选波长的方法(Screening wavelengths based on spectrum ratio analysis,SWSRA),选出仪器间一致性较好且样本间差异大的光谱特征波长,采用筛选出的波长信号建立待测性质的偏最小二乘近红外光谱定标模型。以80个玉米样品中水分、油、蛋白质含量及72个黄芩样品中黄芩苷含量的NIRS预测对该方法进行了检验。结果表明,SWSRA主机模型预测从机样品的各成分含量的平均相对误差均小于4.3%,明显优于全波长模型直接传递的结果,且其预测均方根残差RMSEP与文献报道的其他模型传递方法的结果相当或更优。SWSRA方法具有模型参数少、稳健、简便易行等优点,可以在同类型近红外光谱仪器之间实现模型的无标样传递。     

鱼粉近红外光谱模型传递应用研究

为探讨光栅型与傅里叶变换型近红外分析仪之间模型传递的应用效果,选取国产鱼粉为近红外光谱样本,DS2500F型近红外分析仪为源仪器,MPA型近红外分析仪为目标仪器,采用分段直接校正(PDS)方法实现近红外光谱传递。分别建立水分、粗蛋白质、粗脂肪、蛋氨酸和赖氨酸等组分的预测模型,通过交互验证决定系数(R2cv)、交互验证标准误差(RMSECV)、马氏距离(MD)、系统偏差(Bias)、预测均方根误差(RMSEP)和相对分析误差(RPD)等参数,多维度评估光谱传递后所建预测模型的效果。结果表明,DS2500F仪器的近红外光谱传递到MPA型仪器时,所建国产鱼粉的水分、粗蛋白质、粗脂肪、蛋氨酸、赖氨酸的预测模型与MPA型仪器原始预测模型各参数对比无显著差异,预测效果基本一致,说明国产鱼粉在DS2500F仪器上的近红外光谱通过传递可以替代MPA型仪器的原始光谱,间接实现了模型传递,且具有良好的适用性和共享性,可提高近红外预测模型的应用效率。     

基于小波变换-遗传算法-偏最小二乘的草莓糖度检测研究

采用近红外漫反射光谱分析技术,对草莓糖度进行了无损检测研究。利用便携式近红外光谱仪采集草莓样品在600～1 100 nm波段内的漫反射光谱数据。首先利用小波变换(WT)多分辨率方法对光谱数据进行去噪预处理,然后利用遗传算法(GA)优选特征波长,最后运用偏最小二乘法(PLS)建立草莓糖度的WT-GA-PLS校正模型。该模型校正集的相关系数R_C为0.9395,校正集的均方根误差RMSEC为0.1615,预测集的相关系数R_P为0.9652,预测集的均方根误差EMSEP为0.5042。与全光谱模型(FS-PLS)和小波变换模型(WT-PLS)相比,该模型预测能力更强,稳健性更优。     

近红外光谱分析模型传递简易方法研究

本文在不同时间安装的多台同型号近红外光谱仪上建立推进剂校正模型时,由于推进剂样品数量少且难于保存,新到仪器在建模时常遇到代表性样品数量严重不足.为此,提出将2台波长一致性好的近红外光谱仪器上采集的光谱组成一个混合校正样品光谱集,使用偏最小二乘法(PLS)建立模型的方法.结果表明,在用户缺少专业模型传递软件情况下,该方法...     

小波变换结合多维偏最小二乘方法用于近红外光谱定量分析

将小波变换和多维偏最小二乘法相结合用于近红外光谱定量校正模型的建立.首先将原始光谱进行小波变换分解,得到系列小波细节系数,通过选取一组受外界因素少、信息强的小波系数组成三维光谱阵,然后再采用多维偏最小二乘法建立校正模型.实验结果表明,该方法所建近红外校正模型的预测能力更强,并更具稳健性.     

基于小波系数的近红外光谱局部建模方法与应用研究

局部建模方法使用与预测样本相似的样本建立模型,可解决光谱响应与浓度之间的非线性问题,扩大模型的适用范围,提高预测准确度。采用小波变换进行数据压缩并利用小波系数之间的欧氏距离作为光谱相似性的判据,实现了近红外光谱定量分析的局部建模方法,避免了样本之间的依赖性。将所建立的方法用于烟草样品中氯含量的测定,100次重复计算得到的预测集均方根误差(RMSEP)平均值为0.0665,标准偏差(σ)为0.0045,优于全局建模和基于主成分的局部建模方法。     

Standardization of near infrared spectra measured on multi-instrument

Calibration model transfer is essential for practical applications of near infrared (NIR) spectroscopy because the measurements of the spectra may be performed on different instruments and the difference between the instruments must be corrected. An approach for calibration transfer based on alternating trilinear decomposition (ATLD) algorithm is proposed in this work. From the three-way spectral matrix measured on different instruments, the relative intensity of concentration, spectrum and instrument is obtained using trilinear decomposition. Because the relative intensity of instrument is a reflection of the spectral difference between instruments, the spectra measured on different instruments can be standardized by a correction of the coefficients in the relative intensity. Two NIR datasets of corn and tobacco leaf samples measured with three instruments are used to test the performance of the method. The results show that, for both the datasets, the spectra measured on one instrument can be correctly predicted using the partial least squares (PLS) models built with the spectra measured on the other instruments.     

基于尺度不变特征变换筛选稳定特征波长的近红外光谱模型传递方法

该研究利用一维尺度不变特征变换(SIFT)算法寻找烟叶近红外光谱(Near infrared spectroscopy,NIRS)的稳定特征波长,根据样品精密度测试光谱筛选的波长计算重现率和重现度,采用L_9(3~3)正交表优化SIFT算法中的相关参数,使重现率和重现度尽可能高。基于优化的参数和主机上10个代表性样品的光谱,筛选出10个稳定特征波长集合,以这些波长集合并集的光谱响应为自变量,采用偏最小二乘(PLS)方法构建烟叶总植物碱NIRS模型(简称SIFT-PLS)。该模型直接传递到3台从机后,对3台从机样品总植物碱的平均相对预测误差(MRE)均满足小于6%的企业内控要求,而全光谱模型(WW-PLS)直接转移后仅1台从机的MRE满足要求,经分段直接校正(PDS)方法校正从机光谱后,WW-PLS模型也仅对1台从机的MRE小于6%。采用SIFT算法筛选稳定特征波长建立的NIRS模型可在3台从机直接共享,无需转移集,不需对从机光谱或光谱模型进行校正,实现了真正意义的无标样NIRS模型的直接转移。     

基于近红外光谱的维生素C片无损鉴别分析

基于近红外光谱技术与化学计量学方法,建立了一种国内外不同品牌维生素C片的无损鉴别方法。采集了国内外8个品牌的维生素C片共计40个样本的近红外光谱数据,比较了完整样品以及粉末样品的近红外光谱,采用连续小波变换技术消除背景干扰和基线漂移,基于标准偏差与相对标准偏差的变量筛选方法筛选出具有代表性的波数点,结合主成分分析方法对国内外不同品牌维生素C片进行鉴别分析。结果表明:原始光谱存在着明显的背景干扰和基线漂移现象,且粉末样品的重现性要优于完整样品;单纯使用原始光谱无法辨别来自不同品牌的维生素C片;连续小波变换可以有效消除背景干扰,提高模型鉴别能力;完整样品的鉴别准确率优于粉末样品,说明国内外不同品牌维生素C片主要成分基本一致,可能是辅剂和工艺上存在细微差异。通过结合近红外光谱分析技术与化学计量学方法,可实现对国产以及进口不同品牌维生素C片的鉴别分析。     

最小角回归结合核极限学习机的近红外光谱对柑橘黄龙病的鉴别

传统的柑橘黄龙病检测方法存在准确度低、稳定性差等问题,该文提出了一种基于最小角回归结合核极限学习机(Least angle regression combined with kernel extreme learning machine,LAR-KELM_((RBF)))的近红外柑橘黄龙病鉴别方法。该方法将光谱数据通过小波变换进行预处理,然后用最小角回归(LAR)算法进行光谱波长的筛选,最后通过核极限学习机(KELM_((RBF)))实现样本的分类。实验采用柑橘叶片的近红外光谱数据,验证了LAR-KELM_((RBF))算法的性能,其分类准确度最高为99.91%,标准偏差为0.11。不同规模训练集的实验结果表明,LAR-KELM_((RBF))模型较极限学习机(ELM)、波形叠加极限学习机(SWELM)、反向传播神经网络(BP_((2层)))、KELM_((RBF))和支持向量机(SVM)模型分类准确度高、稳定性强,能够广泛应用于柑橘黄龙病的检测鉴别。     

Multivariate calibration of spectral data using dual-domain regression analysis

To date, few efforts have been made to take simultaneous advantage of the local nature of spectral data in both the time and frequency domains in a single regression model. We describe here the use of a novel chemometrics algorithm using the wavelet transform. We call the algorithm dual-domain regression, as the regression step defines a weighted model in the time-domain based on the contributions of parallel, frequency-domain models made from wavelet coefficients reflecting different scales. In principle, any regression method can be used, and implementation of the algorithm using partial least squares regression and principal component regression are reported here. The performance of the models produced from the algorithm is generally superior to that of regular partial least squares (PLS) or principal component regression (PCR) models applied to data restricted to a single domain. Dual-domain PLS and PCR algorithms are applied to near infrared (NIR) spectral datasets of Cargill corn samples and sets of spectra collected on batch chemical reactions run in different reactors to illustrate the improved robustness of the modeling.     

Ensemble wavelet modelling for determination of wheat and gasoline properties by near and middle infrared spectroscopy

The wavelet transform has been shown to be a useful tool for multivariate calibration. However, the choice of wavelet transform settings (wavelet family, length and number of decomposition levels) for a given application is still an open problem. The present paper proposes an alternative approach, which consists of generating an ensemble model by combining individual models obtained with different wavelet transform settings. The advantages of the proposed method are demonstrated in two analytical problems, namely the determination of moisture and protein in wheat by near infrared spectroscopy and the determination of specific mass and three distillation temperatures (T10, T50, T90) in gasoline by middle infrared spectroscopy. In these problems, the results varied considerably among individual models, which underlines the risk associated to an inadequate choice of wavelet transform settings. In contrast, the ensemble model always provided adequate results in terms of prediction error and noise sensitivity. The proposed method can be seen as an advantageous alternative for multivariate calibration in the wavelet domain, as it frees the analyst from the need to choose a particular configuration for the wavelet transform.     

近红外光谱混合模型定量分析不同物理状态样品的研究

近红外光谱(NIRS)以漫反射模式对非均质样本进行测量时,由于其光谱散射和吸收系数差异较大,建立的校正模型准确性和稳健性较低,因此,本研究提出了一种基于均质样本和模型转移方法建立混合模型的策略,解决非均质样本近红外光谱检测的问题.以烟叶样本为研究对象,分别建立了基于Shenk专利算法(Shenk′s)、分段直接标准化(PDS)和基于典型相关分析的模型转移算法(CTCCA)的烟粉+烟丝、烟粉+烟片混合模型,用于烟丝和烟片样本中烟碱含量的预测.结果表明,混合模型对烟丝和烟片样本的预测均方误差(RMSEP)较直接建模分别降低了1.39%和2.73%,预测结果有一定的改善,稳健性提高,3种方法中CTCCA表现最优.因此,采用近红外光谱均质模型和模型转移方法建立的混合模型对非均质样本的测定具有可行性,有利于在线近红外光谱分析技术的发展,可为近红外光谱模型的共享提供参考.     

Content uniformity determination of pharmaceutical tablets using five near-infrared reflectance spectrometers: a process analytical technology (PAT) approach using robust multivariate calibration transfer algorithms

Near-infrared calibration models were developed for the determination of content uniformity of pharmaceutical tablets containing 29.4% drug load for two dosage strengths (X and Y). Both dosage strengths have a circular geometry and the only difference is the size and weight. Strength X samples weigh approximately 425 mg with a diameter of 12 mm while strength Y samples, weigh approximately 1700 mg with a diameter of 20 mm. Data used in this study were acquired from five NIR instruments manufactured by two different vendors. One of these spectrometers is a dispersive-based NIR system while the other four were Fourier transform (FT) based. The transferability of the optimized partial least-squares (PLS) calibration models developed on the primary instrument (A) located in a research facility was evaluated using spectral data acquired from secondary instruments B, C, D and E. Instruments B and E were located in the same research facility as spectrometer A while instruments C and D were located in a production facility 35 miles away. The same set of tablet samples were used to acquire spectral data from all instruments. This scenario mimics the conventional pharmaceutical technology transfer from research and development to production. Direct cross-instrument prediction without standardization was performed between the primary and each secondary instrument to evaluate the robustness of the primary instrument calibration model. For the strength Y samples, this approach was successful for data acquired on instruments B, C, and D producing root mean square error of prediction (RMSEP) of 1.05, 1.05, and 1.22%, respectively. However for instrument E data, this approach was not successful producing an RMSEP value of 3.40%. A similar deterioration was observed for the strength X samples, with RMSEP values of 2.78, 5.54, 3.40, and 5.78% corresponding to spectral data acquired on instruments B, C, D, and E, respectively. To minimize the effect of instrument variability, calibration transfer techniques such as piecewise direct standardization (PDS) and wavelet hybrid direct standardization (WHDS) were used. The PDS approach, the RMSEP values for strength X samples were lowered to 1.22, 1.12, 1.19, and 1.08% for instruments B, C, D, and E, respectively. Similar improvements were obtained using the WHDS approach with RMSEP values of 1.36, 1.42, 1.36, and 0.98% corresponding to instruments B, C, D, and E, respectively.