Multivariate analysis of fMRI data by oriented partial least squares

Partial least squares (PLS) has been used in multivariate analysis of functional magnetic resonance imaging (fMRI) data as a way of incorporating information about the underlying experimental paradigm. In comparison, principal component analysis (PCA) extracts structure merely by summarizing variance and with no assurance that individual component structures are directly interpretable or that they represent salient and useful features. Oriented partial least squares (OrPLS) is a new PLS-like analysis paradigm in which extracted components can be oriented away from undesirable noise or confounds in the data and toward a desired targeted structure reflecting the fMRI experiment.     

Spatially aggregated multiclass pattern classification in functional MRI using optimally selected functional brain areas

In previous works, boosting aggregation of classifier outputs from discrete brain areas has been demonstrated to reduce dimensionality and improve the robustness and accuracy of functional magnetic resonance imaging (fMRI) classification. However, dimensionality reduction and classification of mixed activation patterns of multiple classes remain challenging. In the present study, the goals were (a) to reduce dimensionality by combining feature reduction at the voxel level and backward elimination of optimally aggregated classifiers at the region level, (b) to compare region selection for spatially aggregated classification using boosting and partial least squares regression methods and (c) to resolve mixed activation patterns using probabilistic prediction of individual tasks. Brain activation maps from interleaved visual, motor, auditory and cognitive tasks were segmented into 144 functional regions. Feature selection reduced the number of feature voxels by more than 50%, leaving 95 regions. The two aggregation approaches further reduced the number of regions to 30, resulting in more than 75% reduction of classification time and misclassification rates of less than 3%. Boosting and partial least squares (PLS) were compared to select the most discriminative and the most task correlated regions, respectively. Successful task prediction in mixed activation patterns was feasible within the first block of task activation in real-time fMRI experiments. This methodology is suitable for sparsifying activation patterns in real-time fMRI and for neurofeedback from distributed networks of brain activation.     

应用近红外光谱技术实现转双价基因(Cry1Ab/Cry2Aj-G10evo)玉米的快速识别

转基因技术在过去的几十年里快速发展, 然而此项技术对生态环境、伦理道德等可能带来的影响尚存争议,因此针对农作物的转基因成分检测和鉴别的相关技术研究十分重要。本研究以转双价基因(cry1Ab/cry2Aj-G10evo)玉米籽粒和玉米面粉为研究对象,采用近红外光谱仪采集900～1 700 nm波段范围的光谱,结合 Savitzky-Golay(SG)平滑算法对提取出的光谱数据进行去除噪声处理。基于全波段光谱和PCA主成分分别建立了偏最小二乘判别分析(PLS)和支持向量机判别模型(SVM)。试验结果表明,在转基因玉米籽粒全谱的判别分析模型中,SVM判别模型效果要优于PLS判别模型,SVM模型识别正确率达到90%以上,PLS的模型识别率只有85%左右。以PCA降维后建立的模型中,SVM模型也取得了最优的效果,建模集和预测集识别正确率达到100%。虽然转基因玉米在研磨加工后外源蛋白和DNA有所下降,但是转基因玉米粉末基于全波段光谱建立的SVM模型的建模集正确率仍有90.625%。结果表明应用近红外光谱技术集合化学计量学方法对转基因玉米的鉴别是可行的,为转基因玉米乃至其他转基因农产品的鉴别提供了技术支持,具有重要的理论意义和应用价值。     

基于NIR分析和模式识别技术的玉米种子识别系统

模式识别技术及数据挖掘方法已成为化学计量学的研究热点。近红外(NIR)光谱分析以其快速、简便、非破坏性等优势广泛应用于光谱信号的处理和分析模型的建立。基于五种不同的模式识别方法:局部线性嵌入(LLE),小波变换(WT),主成分分析(PCA),偏最小二乘(PLS)和支持向量机(SVM),利用NIR技术建立了玉米种子的模式识别系统,并将其应用于108玉米杂交种和母本178种子的近红外光谱样品。首先利用LLE,WT,PCA,PLS进行消噪或降维,然后运用SVM进行分类识别,而一模支持向量机(1-normSVM)算法直接进行分类识别。三个不同NIR光谱范围的数值实验显示:PCA+SVM,LLE+SVM和PLS+SVM识别效果甚佳,而WT+SVM和1-norm SVM方法也有较高的分类精度。实验结果表明了本文提出方法的可行性和有效性,为利用近红外光谱和模式识别技术进行种子识别研究提供了理论依据和实用方法。     

基于NIR分析和模式识别技术的玉米种子识别系统

模式识别技术及数据挖掘方法已成为化学计量学的研究热点。近红外(NIR)光谱分析以其快速、简便、非破坏性等优势广泛应用于光谱信号的处理和分析模型的建立。文章基于五种不同的模式识别方法:局部线性嵌入(LLE),小波变换(WT),主成分分析(PCA),偏最小二乘(PLS)和支持向量机(SVM),利用NIR技术建立了玉米种子的模式识别系统,并将其应用于108玉米杂交种和母本178种子的近红外光谱样品。首先利用LLE,WT,PCA,PLS进行消噪或降维,然后运用SVM进行分类识别,而一模支持向量机(1-norm SVM)算法直接进行分类识别。三个不同NIR光谱范围的数值实验显示:PCA+SVM,LLE+SVM,PLS+SVM识别效果甚佳,而WT+SVM和1-norm SVM方法也有较高的分类精度。实验结果表明了本文提出方法的可行性和有效性,为利用近红外光谱和模式识别技术进行种子识别研究提供了理论依据和实用方法。     

Detection of Apple Marssonina Blotch with PLSR,PCA, and LDA Using Outdoor Hyperspectral Imaging

In this study, hyperspectral images were used to detect a fungal disease in apple leaves called Marssonina blotch(AMB). Estimation models were built to classify healthy, asymptomatic and symptomatic classes using partial least squares regression(PLSR), principal component analysis(PCA), and linear discriminant analysis(LDA) multivariate methods. In general, the LDA estimation model performed the best among the three models in detecting AMB asymptomatic pixels, while all the models were able to detect the symptomatic class. LDA correctly classified asymptomatic pixels and LDA model predicted them with an accuracy of 88.0%. An accuracy of 91.4% was achieved as the total classification accuracy. The results from this work indicate the potential of using the LDA estimation model to identify asymptomatic pixels on leaves infected by AMB.     

Interpretation of FTIR Spectra by Principal Components–Artificial Neural Networks

In order to improve the training speed and increase the predictive ability of artificial neural networks, principal component analysis (PCA) and partial least squares (PLS) were introduced to compress the original data. The principal components (PCs) of FTIR spectroscopic data matrix were obtained by PCA and PLS methods respectively, which were used as the inputs of neural networks. Results indicated that improvement was achieved in three aspects when the PCs instead of the original data were input to the networks. First, iterations were distinctly decreased from 8000 to less than 10. Second, computation time was shortened from 34.95 s to less than 1 s. Third, standard error of prediction (%SEP), mean relative error (MRE), and the root mean square error of prediction (RMSEP) decreased by 35% for the singular value decomposition–artificial neural network (SVD‐ANN) and 80% for the nonlinear iterative partial least squares–ANN (NIPALS‐ANN) or so, which means that the predictive ability was improved significantly. In addition, F‐test was introduced to compare the performance of PCA and PLS for compression of original data, and it was shown that the latter model was more efficient. The presented methodologies of variable selection provide a simple and rapid technique for ANN to interpret FTIR spectra accurately and are advantageous to the widespread use of artificial neural networks.     

Comparison of multivariate linear regression methods in micro‐Raman spectrometric quantitative characterization

Chemical imaging was used in this study as a powerful analytical tool to characterize pharmaceuticals in solid form. The majority of analyses are evaluated with bilinear modelling using only the pure component spectra or just the chemical images themselves to estimate the concentrations in each pixel, which are far from true quantitative determination. Our aim was to create more accurate concentration images using regression methods. For the first time in chemical imaging, variable selections with interval partial least squares (PLS) and with genetic algorithms (PLS‐GA) were applied to increase the efficiency of the models. These were compared to numerous bilinear modelling and multivariate linear regression methods such as univariate regression, classical least squares (CLS), multivariate curve resolution–alternating least squares (MCR‐ALS), principal component regression (PCR) and partial least squares (PLS). Two component spray‐dried pharmaceuticals were used as a model. The paper is shown that, in contrast to the usual way of using either external validation or cross‐validation, both should be performed simultaneously in order to get a clear picture of the prediction errors and to be able to select the appropriate models. Using PLS with variable selection, the root mean square errors were reduced to 3% per pixel by keeping only those peaks that are truly necessary for the estimation of concentrations. It is also shown that interval PLS can point out the best peak for univariate regression, and can thereby be of great help even when regulations allow only univariate models for product quality testing. Variable selection, besides yielding more accurate overall concentrations across a Raman map, also reduces the deviation among pixel concentrations within the images, thereby increasing the sensitivity of homogeneity studies. Copyright © 2015 John Wiley & Sons, Ltd.     

近红外光谱指纹分析在羊肉产地溯源中的应用

为寻求低廉、快速有效地签别羊肉产地来源的方法,对来自内蒙古自治区锡林郭勒盟、呼伦贝尔市和阿拉善盟三个牧区,及重庆市和山东省菏泽市两个农区共99份羊肉样品进行近红外光谱扣描,利用主成分分析结合线性判别分析(PCA+LDA),以及偏最小二乘判别分析法(PLS-DA)对光谱数据进行了分析,建立了羊肉产地来源的定性判别模型.结...     

NIR光谱的Isomap-PLS非线性建模方法

针对样品的近红外(NIR)光谱与其物理化学性质之间存在的非线性关系,提出了一种结合等距映射(Isomap)和偏最小二乘(PLS)的非线性建模新方法。Isomap是一种新的非线性降维方法,属于流形学习方法,能有效地发现高维数据中的本真低维结构。Isomap-PLS建模方法首先用Isomap对高维NIR光谱数据作非线性降维,再用PLS降维并建立校正模型。将Isomap-PLS建模方法分别应用于两个公开的NIR光谱标准数据集,并与PLS单独建模进行比较。结果表明,在两个数据集上,用Isomap-PLS方法建立的校正模型比单独用PLS算法建立的校正模型具有更小的交叉验证均方根误差(RMSECV);对某些性质数据,Isomap-PLS模型比PLS模型的RMSECV值要小2～5倍。因此,Isomap能够有效反映NIR光谱中存在的非线性结构,Isomap-PLS比PLS具有更好的建模与预测能力。     

Megavariate analysis of environmental QSAR data. Part I – A basic framework founded on principal component analysis (PCA), partial least squares (PLS), and statistical molecular design (SMD)

This paper introduces principal component analysis (PCA), partial least squares projections to latent structures (PLS), and statistical molecular design (SMD) as useful tools in deriving multi- and megavariate quantitative structure-activity relationship (QSAR) models. Two QSAR data sets from the fields of environmental toxicology and environmental chemistry are worked out in detail, showing the benefits of PCA, PLS and SMD. PCA is useful when overviewing a data set and exploring relationships among compounds and relationships among variables. PLS is the regression extension of PCA and is used for establishing QSARs. SMD is essential for selecting informative training and test sets of compounds for QSAR calibration and validation.     

近红外光谱技术快速检测莲子粉

莲子是我国重要的药食同源食物,与莲子营养价值相当、便于食用的莲子粉备受消费者青睐。为保证莲子粉的品质,利用近红外光谱(NIRs)技术对掺杂小麦粉、玉米粉和地瓜粉的莲子粉进行鉴定,在样品类别已知下利用支持向量机(SVM)、最小二乘支持向量机(LS-SVM)、偏最小二乘法-判别分析(PLS-DA)模型进行判别,在样品类别未知下基于聚类算法进行判别。同时,对莲子粉中水分含量利用偏最小二乘(PLS)回归进行定量分析。结果表明,LS-SVM模型对纯莲子粉样品与掺入小麦粉、玉米粉和地瓜粉的莲子粉样品的判别率达到100%;基于聚类算法能够有效识别掺入5%地瓜粉、小麦粉和玉米粉的莲子粉样品;PLS模型对莲子粉中水分含量预测综合性能良好,其中经过标准化预处理得到模型效果最佳,其R2_c,RMSEC,R2_p和RMSEP分别达到0.973 2,0.111 5,0.969 5和0.118 9。近红外光谱技术能为隐蔽的莲子粉掺杂的鉴别以及莲子粉中水分含量监控提供一种快速、准确、无损检测的分析方法,为保证高档次莲子品质提供一种有益的思路。     

Quantum partial least squares regression algorithm for multiple correlation problem

Partial least squares (PLS) regression is an important linear regression method that efficiently addresses the multiple correlation problem by combining principal component analysis and multiple regression. In this paper, we present a quantum partial least squares (QPLS) regression algorithm. To solve the high time complexity of the PLS regression, we design a quantum eigenvector search method to speed up principal components and regression parameters construction. Meanwhile, we give a density matrix product method to avoid multiple access to quantum random access memory (QRAM) during building residual matrices. The time and space complexities of the QPLS regression are logarithmic in the independent variable dimension n, the dependent variable dimension w, and the number of variables m. This algorithm achieves exponential speed-ups over the PLS regression on n, m, and w. In addition, the QPLS regression inspires us to explore more potential quantum machine learning applications in future works.     

基于光谱技术和连续投影算法的润滑油品牌快速鉴别研究

为了实现润滑油品牌的快速无损检测,提出了一种基于可见/近红外光谱透射技术与连续投影算法相结合的润滑油品牌快速鉴别新方法。采用连续投影算法对6种润滑油的可见/近红外光谱数据进行波长变量的筛选,再结合偏最小二乘法建立润滑油品牌的鉴别模型。结果表明,鉴别模型的相关系数r为0.9721,预测均方根误差RMSEP为0.4055,鉴别正确率为91.7%。说明提出的连续投影算法结合偏最小二乘算法具有很好的预测效果。     

Fourier‐transform Raman analysis of milk powder: a potential method for rapid quality screening

In this work, FT‐Raman spectroscopy was explored as a fast and reliable screening method for the assessment of milk powder quality and the identification of samples adulterated with whey (1–40% w/w). Raman measurements can easily differentiate milk powders without the need of sample preparation, whereas the traditional methods of quality control, including high‐performance liquid chromatography, are laborious and slow. The FT‐Raman spectra of whole, low‐fat, and skimmed milk powder samples were obtained and distinguished from commercial milk powder samples. In addition, the exploratory analysis employing data from Raman spectroscopy and principal component analysis (PCA)allowed the separation of milk powder samples according to type,identifying differences between samples in the same group. Multivariate analysis was also developed to classify the adulterated milk powder samples using PCA and partial least squares discriminate analysis (PLS‐DA). The resulting PLS‐DA model correctly classified 100% of the adulterated samples. These results clearly demonstrate the utility of FT‐Raman spectroscopy combined with chemometrics as a rapid method for screening milk powder. Copyright © 2011 John Wiley & Sons, Ltd.     

一种用于可见-近红外光谱特征波长选择的新方法

提出了一种基于模拟退火(SA)算法和最小二乘法支持向量机(LS-SVM)选择可见一近红外光谱特征波长的新方法(SA-LS-SVM).该方法用LS-SVM作为识别器,用识别率作为SA的目标函数,提取合适的特征波长数以及对应的特征波长.3种不同品牌的润滑油可见-近红外光谱的特征波长分别用SA_LS-SVM,主成分回归分析(PCA)和偏最小二乘法(PLS)进行处理,提取特征波长或主成分,然后结合反向传播人工神经网络(BP-ANN)对各种处理方法进行识别预测.结果发现,SA-LS-SVM只需从751个数据光谱中提取4个特征波长,就可以使三种品牌润滑油的识别率达到了100%,而其他所有的方法发现预测率都达不到100%,由此验证了SA_LS-SVM的优越性.实验结果表明,SA-LS-SVM不仅能有效地减少建模的变量数,而且可以提高预测精度.     

Identification of waste cooking oil and vegetable oil via Raman spectroscopy

This paper made a qualitative identification of ordinary vegetable oil and waste cooking oil based on Raman spectroscopy. Raman spectra of 73 samples of four varieties oil were acquired through the portable Raman spectrometer. Then, a partial least squares discriminant analysis (PLS‐DA) model and a discrimination model based on characteristic wave band ratio were established. A classification variable model of olive oil, peanut oil, corn oil and waste cooking oil that was established through the PLS‐DA model could identify waste cooking oil accurately from vegetable oils. The identification model established based on selection of waveband characteristics and intensity ratio of different Raman spectrum characteristic peaks could distinguish vegetable oils from waste cooking oil accurately. Research results demonstrated that both ratio method and PLS‐DA could identify waste cooking oil samples accurately. The identification model based on characteristic waveband ratio is simpler than PLS‐DA model. It is widely applicable to identification of waste cooking oil. Copyright © 2016 John Wiley & Sons, Ltd.     

基于高光谱成像技术的鲜枣裂纹的识别研究

裂纹是衡量鲜枣品质的重要指标之一,果皮裂纹加速鲜枣的腐烂,导致鲜枣货架期的缩短,严重降低鲜枣的经济价值。采用高光谱成像技术在380～1 030 nm波段范围内对鲜枣裂纹的位置及大小信息特征进行快速识别。选用偏最小二乘回归(PLSR)、连续投影法(SPA)和全波段图像主成分分析(PCA),得到鲜枣裂纹相关的敏感波段。然后利用选取的鲜枣裂纹的敏感波段对建模集的132个样本建立最小二乘支持向量机(LS-SVM)判别模型,并对预测集的44个样本进行判别。对PLSR-LS-SVM,SPA-LS-SVM和PCA-LS-SVM判别模型采用ROC曲线进行评判,得出PLSR-LS-SVM模型对鲜枣裂纹定性判别的结果(area=1,std=0)最佳。选取PLSR回归系数挑选出的5条鲜枣裂纹敏感波段(467,544,639,673和682 nm)对应的单波段图像进行主成分分析,其中将主成分PC4的图像结合图像处理技术,最终识别出鲜枣裂纹的位置、大小信息。结果表明,采用高光谱成像技术结合光谱图像处理可以实现鲜枣裂纹定性判别和定量识别的研究,为进一步开发相关仪器的研究提供理论方法和依据。     

基于近红外光谱技术构建蓝莓总酚含量的模型

选用6个品种(埃利奥特、达柔、爱国者、杜克、北蓝、蓝丰)的蓝莓,应用傅里叶变换近红外光谱仪(FTNIR)对蓝莓中总酚含量进行分析,采用主成分分析(PCA)和偏最小二乘回归法(PLS)建立了蓝莓总酚含量近红外数学校正集模型,其相关系数为0.9512、校正集标准偏差(RMSEC)为0.72、预测集标准偏差(RMSEP)为0...     

基于近红外光谱和偏最小二乘法的慈竹纤维素结晶度预测研究

将近红外光谱技术和化学计量学相结合分析慈竹纤维素结晶度。通过区间偏最小二乘法(iPLS)、联合区间偏最小二乘法(siPLS)和反向区间偏最小二乘法(biPLS)优化建模区域,建立经多元散射校正后光谱的结晶度分析模型,并与全光谱范围350～2 500nm建立的偏最小二乘(PLS)模型进行比较。结果表明,三种改进偏最小二乘法建立的结晶度模型预测效果均优于PLS模型,并且当采用联合区间偏最小二乘法将全光谱进行30个子区间划分,选择三个子区间[8 12 19]组合时,建立的siPLS模型预测效果最好,相关系数(r)达到0.88,预测标准差(RMSEP)为0.0117。因此,采用联合区间偏最小二乘法可以有效选择建模光谱区域,提高模型预测能力,实现慈竹纤维素结晶度的快速预测。