高效液相色谱-电化学检测指纹图谱鉴别3种单花种蜂蜜花源

基于高效液相色谱-电化学检测器(high performance liquid chromatography-electrochemical detection,HPLC-ECD)技术建立一种新的蜂蜜花源鉴别方法。以采自中国不同地区的3种单花种蜂蜜为研究对象,构建了3种单花种蜂蜜的HPLC-ECD指纹图谱,提取HPLC-ECD图谱共有峰面积信息并应用主成分分析和系统聚类分析进行蜂蜜花源分类,并对完全未参与建模的蜂蜜样品进行验证。结果表明,45个蜂蜜样品(枸杞蜜、荆条蜜、荔枝蜜各15个),均可通过主成分分析和系统聚类分析按照其花源正确分类,正确率达到100%。该蜂蜜花源鉴别的模型对完全未参与建模的枸杞蜜、荆条蜜和荔枝蜜样品的正确预判率可达到100%、80%和100%。研究表明,HPLCECD指纹图谱技术应用主成分分析和系统聚类分析可以作为一种快速、准确、绿色的判别蜂蜜花源的方法。     

气相离子迁移谱技术鉴别重庆三峡库区特色中蜂蜜研究

采用气相离子迁移谱(gas chromatography-ion mobility spectrometry, GC-IMS)技术对来源于重庆三峡库区的油菜花、五倍子花、枇杷花和柑橘花的4种特色中华蜜蜂蜂蜜(中蜂蜜)的挥发性有机成分进行测定和分析,建立不同植物来源中蜂蜜的判别模型,并对不同植物来源蜂蜜进行鉴别和分类。利用二维差谱法筛选出58个有效特征成分的特征峰作为表征中蜂蜜植物来源差异信息的特征变量,利用主成分分析(principal component analysis, PCA)和线性判别分析(linear discriminate analysis, LDA)方法建立判别模型。结果表明,该研究选取的特征变量经PCA处理后前3个主成分的累积贡献率为81.35%,得分图中4种中蜂蜜分布于不同的区域,无重叠或交叉,建立的判别模型可有效识别不同植物来源的中蜂蜜,准确率为100%,对其他植物来源中蜂蜜的误判率仅为5.9%。该研究利用GC-IMS测定中蜂蜜中的挥发性有机成分,结合PCA和LDA可以准确地区分重庆三峡库区不同蜜源植物特色的中蜂蜜样品,为中蜂蜜的品种鉴别和质量控制提供了新的技术...     

近红外光谱技术定性鉴别蜂蜜品种及真伪的研究

提出了蜂蜜品种及真伪定性鉴别的新方法。在12000～4000cm-1范采集荆条蜜、槐花蜜、油菜蜜和掺假蜜的近红外光谱,结合一阶导、多元散射校正及变量标准化)三种方法对光谱进行预处理,以主成分分析结合马氏距离判别法,在不同谱区建立蜂蜜品种及真伪定性鉴别模型。研究发现6100～5700cm-1谱区为最佳建模波段,品种判别正确率达90%以上,真伪鉴别正确率为93.10%。     

小麦霉菌污染支持向量机判别模型的建立

为建立一种快速判别小麦霉菌污染的方法,该研究采用近红外光谱技术结合化学计量学方法,以126份小麦样品为研究对象,通过剔除异常样品、光谱降维和预处理,采用支持向量机分类（support vector machine classification,SVM）方法建立判别模型。结果表明：运用基于马氏距离的主成分分析方法剔除异常样品5个,将原始光谱数据进行降维处理得到8个主成分,能够代表原始样本的98.80%。输入变量的最佳预处理方式为标准正态变量变换,最佳核函数为linear,核函数参数C值为10,SVM判别模型的训练集判别正确率为100%,交叉验证判别正确率为98.89%。用未参与建立判别模型的外部验证集样品对SVM判别模型进行验证,结果表明：SVM判别模型对外部验证集样品的判别正确率为100%。该研究所建立的SVM判别模型可以用于小麦霉菌污染的快速检测。     

核磁共振技术结合化学计量学分析不同来源单花蜜的差异

采用核磁共振(NMR)技术结合化学计量学分析方法对4种蜜源的单花蜜(洋槐蜜、荆条蜜、葵花蜜和麦卢卡蜜)的核磁共振氢谱进行差异分析。优化前处理方法并对数据进行预处理后,采用无监督的主成分分析(PCA)和有监督的偏最小二乘判别分析(PLS-DA)、正交偏最小二乘判别分析(OPLS-DA)等多元统计分析方法从核磁信号中提取各组的分类信息。结果表明:采用分散液液萃取法(DLLME)可避免蜂蜜中大量糖对低含量特征性强的化合物的信号掩盖,更好地涵盖δ0.5～10.0范围所有区域信号;采用log转换和帕莱托换算方式预处理数据,PCA模型可以很好地区分4种蜂蜜;利用PLS-DA所建模型对4种蜂蜜的判别解释能力达93.2%,对未知样本的预测能力为87.6%;采用置换测试法的交叉验证表明模型可靠且稳健;基于两两组间OPLS-DA模型可识别4种蜂蜜相互区分的相应核磁分段位移作为特征性变量。该方法简单、快速,可以扩展到多种蜜源单花蜜的区分、识别,为不同单花蜜的质量评价建立有效方法和预测模式。     

傅里叶变换红外光谱鉴别芝麻酱中掺杂花生酱

提出了基于傅里叶变换红外光谱鉴别纯芝麻酱中掺杂花生酱的方法。使用傅里叶变换红外光谱仪采集650~4000cm-1范围内的纯芝麻酱样品和掺杂花生酱的芝麻酱样品共计93例,以标准正态变量变换(standard normal variate,SNV)做光谱预处理,比较了支持向量机(support vector machine,SVM)、线性判别分析(liner discriminant analysis,LDA)2种判定方法,并选择有效波长建立模型。研究表明,采用SVM建立判别模型较优;选取3071~2792cm-1,1786~667cm-1为有效波长后SVM模型的综合判别正确率提高至93.55%,且模型输入变量减少58.20%。采用傅里叶变换红外光谱结合支持向量机判别模型可以较好地应用于纯芝麻酱中掺杂花生酱的鉴别。     

可见/近红外结合MIA变量优选和支持向量机判别山茶油的制取方式

山茶油的主要制取方式有压榨法和浸出法,且压榨山茶油的品质优于浸出法。本研究利用可见/近红外光谱技术结合化学计量学对山茶油的制取方式进行判别研究。采集不同制取方式的山茶油样本在350~1800nm波段范围的可见/近红外光谱,利用边界影响分析(margin influence analysis,MIA)新方法进行波长变量优选,并应用支持向量机(support vector machines,SVM)对优选的波长变量建立山茶油制取方式的判别分类模型。结果表明:可见/近红外光谱联合MIA-SVM方法判别山茶油的制取方式是可行的,其校正集和预测集样本的灵敏度、特异性及正确率分别为100%、87.50%、93.75和100%、87.50%、93.75%。说明MIA是一种有效的波长变量选择方法,能简化分类模型,提高分类模型的稳定性和预测精度。     

基于矿物元素与同位素联合分析的吉林大米产地判别的研究

为实现更加准确、可靠的吉林省大米产地判别分析,利用电感耦合等离子质谱仪(ICP-MS)和同位素质谱仪(IRMS)测定样品中的12种矿物元素和δ¹³C、δ¹⁵N同位素含量,结合线性判别分析(LDA)和支持向量机(SVM)模型探究同位素数据的加入对产地确证结果的影响;利用Xgboost模型和交叉验证,筛选判别因子最优合集。结果表明,在12种矿物元素数据联合δ¹³C、δ¹⁵N数据分别代入模型后,线性判别分析结果准确率和回代正确率从79%、73.3%提升到80.1%、74.0%;支持向量机模型结果的准确度和精确度从83.019%、82.589%提高到90.245%、87.549%。得到吉林省大米最优判别因子合集为{Si、Mn、Ca、Mg、Al、Na、Cr、P、δ¹³C},为快速准确实现吉林省大米产地溯源提供理论参考。     

基于多源光谱分析技术的鱼油品牌判别方法研究

多源光谱分析技术被用于鱼油品牌快速无损鉴别。采用可见光谱分析技术、短波近红外光谱分析技术、长波近红外光谱分析技术、中红外光谱分析技术和核磁共振光谱分析技术采集了7种不同品牌的鱼油的光谱特征,并应用偏最小二乘判别分析法(partial least squares discrimination analysis,PLS-DA)和最小二乘支持向量机(least-squares support vector machine,LS-SVM)建立判别模型并比较判别结果。基于长波近红外光谱的PLS-DA模型和LS-SVM模型取得了最高识别正确率,建模集和预测集识别正确率均达到100%。采用中红外光谱和核磁共振谱分别建立的LS-SVM模型,也可以获得100%的判别正确率。而可见光谱和短波近红外光谱则判别准确率较差。且LS-SVM算法较PLS-DA更加适合用于建立光谱数据和鱼油品牌之间的判别模型。研究结果表面长波近红外光谱技术能够有效判别不同鱼油的品牌,为将来鱼油品质鉴定便携式仪器的开发提供了技术支持和理论依据。     

基于拉曼光谱和支持向量机的黄酒品质快速分析

本研究使用拉曼光谱分析技术采集不同产地和不同酒龄的黄酒样品指纹信息,对比判别分析(DA)和最小二乘支持向量机(LS-SVM)所建黄酒品质快速模型性能,确定最优模型以实现快速准确地评价黄酒品质。本研究在全波段范围利用主成分分析对拉曼光谱数据降维,计算降维谱图间马氏距离,基于ward’s算法建立判别分析模型;将全波段范围作为最小二乘支持向量机的输入量,选择出能较好处理非线性问题的RBF为核函数,同时采用交互验证方式优化RBF核函数参数,基于优化RBF核函数,建立最小二乘支持向量机鉴别模型。研究结果表明:拉曼光谱结合最小二乘支持向量机鉴别模型对黄酒产地和酒龄的鉴别正确率均为100%;拉曼光谱结合判别分析鉴别模型对嘉善、绍兴和上海黄酒的鉴别正确率分别为100%、80%和80%,对黄酒酒龄的鉴别正确率均为100%;最小二乘支持向量机模型性能优于判别分析模型。拉曼光谱结合化学计量学方法可快速、准确评价黄酒品质。     

Detection of adulterants such as sweeteners materials in honey using near-infrared spectroscopy and chemometrics

Near-infrared (NIR) spectroscopy combined with chemometrics methods has been used to detect adulteration of honey samples. The sample set contained 135 spectra of authentic (n = 68) and adulterated (n = 67) honey samples. Spectral data were compressed using wavelet transformation (WT) and principal component analysis (PCA), respectively. In this paper, five classification modeling methods including least square support vector machine (LS-SVM), support vector machine (SVM), back propagation artificial neural network (BP-ANN), linear discriminant analysis (LDA), and K-nearest neighbors (KNN) were adopted to correctly classify pure and adulterated honey samples. WT proved more effective than PCA, as a means for variables selection. Best classification models were achieved with LS-SVM. A total accuracy of 95.1% and the area under the receiver operating characteristic curves (AUC) of 0.952 for test set were obtained by LS-SVM. The results showed that WT-LS-SVM can be as a rapid screening technique for detection of this type of honey adulteration with good accuracy and better generalization.     

Origin Identification and Quantitative Analysis of Honeys by Nuclear Magnetic Resonance and Chemometric Techniques

The combination of ¹H NMR spectroscopy and multivariate statistical analysis has become a promising method for the discrimination of food origins. In this paper, this method has been successfully employed to analyze 70 Chinese honey samples from eight botanic origins, three geographical origins, and five production dates. Thirty-three components in honey samples were detected and identified from their ¹H NMR spectra, and 20 of them were accurately quantified by comparing their integral area with that of internal standards with relaxation time correction. Nontargeted principal component analysis (PCA) has been applied to distinguish the honeys from different botanical and geographical origins. The variations of components in the honeys, including saccharides and all kind of amino and organic carboxylic acids, confirmed their clustering according to their origins in PCA scores plots. Orthogonal partial least squares discriminant analysis (OPLS-DA) based on the NMR data for the different pairwise honey samples allows to identify the compositional variations contributed to geographical discrimination and storage time. Hence, NMR spectroscopy coupled with chemometric techniques offers an efficient tool for quality control of honey, and it could further serve to the classification, qualitative and quantitative control of other foods.     

基于支持向量机的食醋总酸近红外光谱建模

为了得到稳定可靠的食醋总酸光谱模型,以不同产地、不同种类的95个食醋样品为研究对象,应用基于统计学原理的最小二乘支持向量机(LS-SVM)对食醋总酸含量进行光谱分析.对预处理后的光谱进行主成分分析(PCA),以主成分信号作为输入变量建立食醋总酸含量的近红外光谱模型,并与偏最小二乘法(PLS)和向后区间偏最小二乘法(biPLS)模型进行比较.结果表明,LS-SVM模型中的校正集中的相关系数(rc)和交互验证均方根误差(RMSECV)分别达到0.9614和0.2192,预测集相关系数(rp)和预测均方根误差(RMSEP)分别达到和0.919和0.3226,均优于PLS和biPLS模型.研究表明,近红外光谱与食醋总酸含量呈非线性关系,采用LS-SVM建立的模型预测性能更好,精度更高.     

Geographical Authentication of Tequila According to its Mineral Content by Means of Support Vector Machines

The elemental profile of tequila samples from the three main production areas of the State of Jalisco, namely Amatitlan, Guadalajara, and Tequila, was used to carry out their geographical characterization. With this aim, the concentration of Al, Ba, Ca, Cu, Fe, K, Mg, Mn, Na, S, Sr, and Zn was determined by inductively coupled plasma atomic emission spectroscopy. Principal component analysis was addressed to visualize data trends. The number of input variables was reduced by means of backward stepwise linear discriminant analysis and support vector machines were used to construct an adequate classification model. The best classification performance was obtained by a linear support vector machine model with 100% of prediction ability.     

C/C pattern of honey from Turkey and determination of adulteration in commercially available honey samples using EA-IRMS

Honey adulteration particularly by adding cheap sugars such as High Fructose Corn Syrup (HFCS) and cane sugar syrup into natural honey is widespread. This study reports (¹³C/¹²C, ‰) patterns of 31 authentic honey samples obtained from different sources and regions of Turkey as well as 43 commercial honey samples to determine the adulteration using mass spectrometer coupled to elemental analyzer (EA-IRMS). The analyses indicated that the ranges of (¹³C/¹²C, ‰) values of honey and protein fractions of Turkish honey are from −23.30 to −27.58‰ and −24.13 to −26.76‰, respectively. These values for commercial honey samples were determined to range from −11.28 to −25.54‰ and −19.35 to −25.61‰, respectively. The numbers of adulterated commercial honey samples were found to be 10, which corresponds to 23% of the total number of the samples. Diastase activity, HMF content, electrical conductivity and moisture content of honey samples were also determined. Method validation and uncertainty budget calculations were also reported.     

Characterization and Discrimination of the Floral Origin of Sidr Honey by Physicochemical Data Combined with Multivariate Analysis

Sidr honey represents one of the most expensive monofloral honeys worldwide. The quality control of such honey types usually depends on pollen analysis or comparison of physicochemical characters. In the presented work, 38 different honey samples of which 13 represented genuine Sidr (Ziziphus spina-christy) honey samples were collected from various areas of Yemen. All samples were characterized by physicochemical parameters including moisture content, pH, electrical conductivity, and free acidity. The physicochemical data was subjected to multivariate data analysis including principal component analysis (PCA) and hierarchical cluster analysis (HCA). The development of partial least square discriminant analysis (PLS-DA) model on validation gave 100 % correct classification of the test set samples. All tested honey samples were within the level permitted by the international standards for honey quality. The application of the discriminant technique PLS-DA presented excellent potential for discriminating the botanical origin of Yemeni Sidr honey from other non-Sidr samples and may serve as a discriminant model to be applied to other honey types worldwide.     

Characterization of artisanal honeys from Madrid (Central Spain) on the basis of their melissopalynological,physicochemical and volatile composition data

Forty-six artisanal honey samples, from different places of Madrid province (Central Spain), were characterized on the basis of their melissopalynological, physico-chemical and volatile composition data. Results were submitted to principal component analysis and stepwise discriminant analysis in order to evaluate the existence of data patterns and the possibility of differentiation of Madrid honey samples according to their botanical source (honeydew, nectar) and geographical collection place (mountain, plain). Colour, electrical conductivity, acidity, ash content and pH were the physicochemical parameters with higher discrimination power in the differentiation of nectar and honeydew honeys while, among the volatile components, concentrations of borneol, 1-(2-furanyl)-ethanone and 3-hydroxy-2-butanone were the most discriminant variables. In the differentiation of honey samples from mountain and plain zones, 2,3-butanediol and 1-(2-furanyl)-ethanone were the most significant volatiles, while physicochemical data were not useful for distinguishing between collection places. The honeydew percentage in a honey sample (HD%) was estimated from physicochemical measures and also from volatile concentrations; 2,3-butanediol, 3-hydroxy-2-butanone, 1-hydroxy-2-propanone and 1-(2-furanyl)-ethanone were found to be related to HD%.     

Characterisation of Serbian unifloral honeys according to their physicochemical parameters

The characterisation of three unifloral Serbian honeys (acacia, sunflower and linden) was carried out based on some common physicochemical parameters (water content, electrical conductivity, free acidity, optical rotation and pH). A total of 201 honey samples, collected during the 2009 harvesting season, were analysed. Principal component analysis (PCA) and linear discriminant analysis (LDA) were used to highlight the data structure and to find the relationships between the physicochemical parameters and the botanical origin of honey. The variables that best discriminated the samples were electrical conductivity (ranging from 0.10 to 0.76 mS/cm), free acidity (ranging from 7.80 to 42.70 meq/kg) and pH (ranging from 3.17 to 5.85). LDA resulted in a classification model with a high predictive power, allowing further assessment of unknown samples of the three unifloral honeys. Determination of geographic origin of acacia honey samples based on physicochemical properties and chemometrics was attempted.     

Geographical characterisation of honeys according to their mineral content and antioxidant activity using a chemometric approach

In this article, discrimination models are presented, relating the origin of honey samples to several variables, being the concentrations of different cations and anions in the honey samples measured by ion chromatography, and parameters that measure/reflect the antioxidant activity of the honey samples. The unsupervised method, principal component analysis, and supervised discrimination methods, such as linear and quadratic discriminant analysis, and classification and regression trees (CART), were applied to evaluate the existence of data patterns and the relationship between geographical origin and the measured parameters. The model with the best predictive ability (%CCR_TEST = 66.67%), the best overall % specificity (80%) and the best overall % sensitivity (67%) was found to be CART. It was proven that the mineral content and parameters analysed can provide enough information for the geographical characterisation and discrimination of honey.     

Classification of Brazilian vinegars according to their H NMR spectra by pattern recognition analysis

This work describes using ¹H NMR data and pattern recognition analysis to classify vinegars. Vinegar authenticity is linked to raw ingredient source and manufacturing conditions. Application of PCA and HCA methods resulted in the natural clustering of the samples according to the raw material used. Wine vinegars were characterized by a high concentration of ethyl acetate, glycerol, methanol and tartaric acid, while glycerol and ethyl acetate signals were not visible in alcohol/agrin vinegars. Apple vinegars showed to be richer in alanine. The KNN, SIMCA and PLS-DA methods were used to build predictive models for classification of vinegar type wine, apple and alcohol/agrin (27 samples - 22 as training set). The models were tested using an independent set (5 samples), no samples were wrongly classified. Validated models were used to predict the class of 21 commercial samples, which, as expected, were correctly classified. Eight commercial vinegars (honey, orange, pineapple and rice) were discriminated from these samples using PCA method. Honey vinegars did not present ethanol signals and pineapple vinegars presented the largest amount of tartaric acid. Rice and orange vinegars are richer in lactic acid and did not present the methanol signal. Alanine signals were not visible in orange vinegars.