Rapid classification of virgin and recycled EPS containers by Fourier transform infrared spectroscopy and chemometrics

ABSTRACT

A rapid and sensitive method for classification of virgin and recycled expanded polystyrene (EPS) food containers was developed using Fourier transform infrared spectroscopy (FTIR) and chemometrics. This method includes preparing a transparent film by dissolution, examining by FTIR and developing classification models. The degradation of EPS containers occurring during the recycling process was reflected by the carbonyl region of the infrared spectrum which was used as variables for multivariate data analysis. PCA was used to reduce the data dimension and view the sample similarities. Soft independent modelling of class analogy (SIMCA), partial least squares-discrimination analysis (PLS-DA) and linear discrimination analysis (LDA) were applied to construct three classification models. The best discrimination results were obtained by an LDA model, with all samples correctly classified. PLS-DA and SIMCA could not classify the recycled EPS samples with low levels of adulteration. When applying this method to commercially available EPS containers, about 45% of samples were shown to contain recycled polystyrene resins. It is concluded that the carbonyl region of the infrared spectra coupled with chemometrics could be a powerful tool for the classification of virgin and recycled EPS food containers.     

Simultaneous Classification of Teas According to Their Varieties and Geographical Origins by Using NIR Spectroscopy and SPA-LDA

Due to the human health benefits already scientifically proven, tea (Camellia sinensis) has been widely studied in the literature. Several studies report the classification of the variety or geographical origin of teas, separately. Thus, this paper has proposed a methodology for simultaneous classification of tea samples according to their varieties (green or black) and geographical origins (Brazil, Argentina, or Sri Lanka). For this purpose, near-infrared (NIR) spectroscopy and three differing supervised pattern recognition techniques, namely SIMCA (soft independent modeling of class analogy), PLS-DA (partial least squares-discriminant analysis), and SPA-LDA (successive projections algorithm associated with linear discriminant analysis) have been used. Despite having good results, both full-spectrum PLS-DA and SIMCA were not able to achieve 100 % classification accuracy, regardless of the significance level for the F test in the case of the SIMCA model. On the other hand, the resulting SPA-LDA model successfully classified all studied samples into five differing tea classes (Argentinean green tea; Brazilian green tea; Argentinean black tea; Brazilian black tea; and Sri Lankan black tea) using 12 wave numbers alone.     

Classification of longan fruit bruising using visible spectroscopy

This research showed the potential of using visible spectroscopy for classification of non-bruised and bruised longan fruits. The visible spectra of bruised and non-bruised longan fruits were acquired from 400 to 700 nm with 10 nm resolution by the spectrophotometer. The principal component analysis (PCA), Partial Least Square Discriminant Analysis (PLS-DA) and Soft Independent Modeling of Class Analogy (SIMCA) were used to develop classification models. The Partial Least Square Discriminant Analogy (PLS-DA) showed better classification accuracy than SIMCA with 100% correctness. The result was found to be helpful for the application in the industry for on-line and portable application.     

Identification of components of Brazilian honey by 1H NMR and classification of its botanical origin by chemometric methods

The potential of NMR spectroscopy to differentiate honeys concerning to the nectar employed in its production was evaluated. The application of chemometric methods to ¹H NMR spectra has allowed to discriminate the honeys produced in the state of São Paulo, being identified the signals of responsible substances for the discrimination. Application of PCA and HCA methods to ¹H NMR data have resulted in the natural clustering of the samples. Wildflower honeys were characterized by higher concentration of phenylalanine and tyrosine. Citrus honeys showed higher amounts of sucrose than other compounds, while eucalyptus honeys had higher amount of lactic acid than the others. Assa-peixe honeys showed spectra similar to eucalyptus and citrus. Sugar-cane honeys showed some signals similar to eucalyptus and citrus honeys, but also showed the tyrosine and phenylalanine signals. Adulterated honeys showed 5-hydroxymethylfurfural, citric acid and ethanol signals. KNN, SIMCA and PLS-DA methods were used to build predictive models for honey classification. In the commercial honeys prediction KNN, SIMCA and PLS-DA models correctly classified 66.7; 22.2 and 72.2% of the samples, respectively.     

Simultaneous detection for adulterations of maltodextrin,sodium carbonate,and whey in raw milk using Raman spectroscopy and chemometrics

To achieve rapid on-site identification of raw milk adulteration and simultaneously quantify the levels of various adulterants, we combined Raman spectroscopy with chemometrics to detect 3 of the most common adulterants. Raw milk was artificially adulterated with maltodextrin (0.5–15.0%; wt/wt), sodium carbonate (10–100 mg/kg), or whey (1.0–20.0%; wt/wt). Partial least square discriminant analysis (PLS-DA) classification and a partial least square (PLS) regression model were established using Raman spectra of 144 samples, among which 108 samples were used for training and 36 were used for validation. A model with excellent performance was obtained by spectral preprocessing with first derivative, and variable selection optimization with variable importance in the projection. The classification accuracy of the PLS-DA model was 95.83% for maltodextrin, 100% for sodium carbonate, 95.84% for whey, and 92.25% for pure raw milk. The PLS model had a detection limit of 1.46% for maltodextrin, 4.38 mg/kg for sodium carbonate, and 2.64% for whey. These results suggested that Raman spectroscopy combined with PLS-DA and PLS model can rapidly and efficiently detect adulterants of maltodextrin, sodium carbonate, and whey in raw milk.     

Confirmation of brand identity in foods by near infrared transflectance spectroscopy using classification and class-modelling chemometric techniques — The example of a Belgian beer

Food and beverage processors require tools to monitor conformance of finished goods to their defined specification; regulatory authorities need appropriate methods for detecting retail fraud. In this report, samples (n = 275) of Belgian and other European beers were collected and analysed using near infrared transflectance spectroscopy; three class-modelling techniques (soft independent modelling of class analogy, SIMCA; potential functions techniques, POTFUN; and unequal dispersed classes, UNEQ) were employed to characterise beer types (firstly Trappist and then Rochefort) while a classification method (partial least squares discriminant analysis, PLS-DA) was applied to discriminate between two final beer classes: Rochefort 8° and Rochefort 10°. The class-models and the classification rules developed were validated by means of an external prediction set. A discussion on the appropriate use of these chemometric approaches is included. Modelling of Trappist beers met with limited success while model efficiencies for Rochefort samples were highest for SIMCA and UNEQ applications i.e. 81.4% and 84.5% respectively. The classification of beers as Rochefort 8? or Rochefort 10? was possible with an average correct classification rate of 93.4%.     

Feasibility of conventional and Roundup Ready® soybeans discrimination by different near infrared reflectance technologies

Identification and proper labelling of genetically modified organisms is required and increasingly demanded by legislation and consumers worldwide. In this study, the feasibility of three near infrared reflectance technologies (a chemical imaging unit, a commercial diode array instrument, and a light tube non-commercial instrument) were compared for discriminating Roundup Ready® and not genetically modified soybean seeds. Over 200 seeds of each class (Roundup Ready® and conventional) were used. Principal Component Analysis with Artificial Neural Networks (PCA–ANN) and Locally Weighted Principal Component Regression (LW-PCR) were used for creating the discrimination models. Discrimination accuracies when new tested seeds belonged to samples included in the training sets achieved accuracies over 90% of correctly classified seeds for LW-PCR models. The light tube performed the best, while the imaging unit showed the worse accuracies overall. Models validated with new seeds from samples not included in the training set had accuracies of 72–79%.     

基于连续投影算法的油菜蜜近红外光谱真伪鉴别的研究

采用连续投影算法(successive project algorithm,SPA)对177个不同产地油菜蜜样本的近红外光谱做波长选择,然后以33个特征变量作线性识别分析(LDA)。同时,也采用了主成分分析(PCA)对变量进行压缩。比较了二次识别分析(QDA)和簇类独立软模式分类法(SIMCA)的鉴别结果。SPA-LDA模型预测集的鉴别准确率为97.7%,而PCA-LDA、全谱的SIMCA和SPA-QDA预测集的正确率分别为93.2%、95.4%和90.9%;上述四种方法ROC曲线下的面积分别为0.964、0.912、0.948和0.933。SPA-LDA性能比其他三种方法要好。该方法准确、可靠,为蜂蜜真实性的现场快速检测提供了一种新方法。     

Classification and Authentication of Barley (<Emphasis Type="Italic">Hordeum vulgare</Emphasis>) Malt Varieties: Combining Attenuated Total Reflectance Mid-infrared Spectroscopy with Chemometrics

Attenuated total reflectance (ATR) mid-infrared (MIR) spectroscopy combined with chemometrics was explored as a tool to classify and authenticate Australian barley varieties. Grain samples (n = 162) were sourced from eight commercial barley varieties and analysed in the MIR range. Principal component analysis (PCA), discriminant partial least squares regression (PLS-DA), linear discriminant analysis (LDA) and soft independent modelling of class analogy (SIMCA) were used to classify the barley grain samples according to variety. PLS-DA correctly classified barley varieties between 91 and 100 %. The results have demonstrated the usefulness of ATR-MIR spectroscopy combined with chemometrics as a rapid method to classify barley grain samples according to their variety. Although MIR is not routinely available at the receival point in most of the cereal trade companies, it has the potential to be used in breeding programmes.     

Tracing the geographical origin of honeys based on volatile compounds profiles assessment using pattern recognition techniques

The goal of this study was to examine the possibility of verifying the geographical origin of honeys based on the profiles of volatile compounds. A head-space solid phase microextraction (SPME) combined with comprehensive two-dimensional gas chromatography–time-of-flight mass spectrometry (GC × GC–TOFMS) was used to analyze the volatiles in honeys with various geographical and floral origins. Once the analytical data were collected, supervised pattern recognition techniques were applied to construct classification/discrimination rules to predict the origin of samples on the basis of their profiles of volatile compounds. Specifically, linear discriminant analysis (LDA), soft independent modeling of class analogies (SIMCA), discriminant partial least squares (DPLS) and support vector machines (SVM) with the recently proposed Pearson VII universal kernel (PUK) were used in our study to discriminate between Corsican and non-Corsican honeys. Although DPLS and LDA provided models with high sensitivities and specificities, the best performance was achieved by the SVM using PUK. The results of this study demonstrated that GC × GC–TOFMS combined with methods like LDA, DPLS and SVM can be successfully applied to detect mislabeling of Corsican honeys.     

Lychee Variety Discrimination by Hyperspectral Imaging Coupled with Multivariate Classification

This current study was carried out to investigate the ability of hyperspectral imaging (HSI) technique and multivariate classification for the differentiation of lychee varieties. A total of 122 lychee samples from three varieties (“Baila,” “Jizhui,” and “Guiwei”) were used. The relationship between reflectance spectra and lychee varieties were established. Principal component analysis (PCA) was implemented on the region of interest (ROI) image to reduce data dimensionality and visualize the cluster trend. The first two principal components (PCs) explained over 97 % of variances of all spectral bands. Linear (soft independent modeling of class analogy (SIMCA) and partial least square discriminant analysis (PLS-DA)) and nonlinear (back propagation neural network (BPNN) and support vector machine (SVM)) multivariate classification methods were used to develop discrimination models. The results revealed that SVM model achieved the best result, with the identification rate of 100 % in the calibration set and 87.81 % in the prediction set. BPNN had a discrimination rate of 100 % for the training set and 85.37 % for prediction set, while PSL-DA and SIMCA model had a discrimination rate of 78.05 % and 60.98 % for prediction sets, respectively. The nonlinear classification methods used were superior to the linear ones. The overall results showed that HSI system with SVM classification tool could be used in identification of lychee varieties.     

基于近红外技术和偏最小二乘判别法对无花果成熟度的快速判别

以无花果为试验对象,对其进行近红外光谱采集,并对其糖度、单果重、纵径、横径、硬度5个指标进行K-均值聚类;根据光谱数据、主成分分析确定最优聚类效果的成分和各类别的指标分布构建偏最小二乘判别分析(PLS-DA)模型进行聚类判别,以实现对果实成熟度(幼果期、成长期、成熟期)分类的准确、快速、无损伤鉴别。结果表明,3种成熟阶段的无花果样品的糖度、单果重和硬度均具有显著性差异,成熟果和成长果与幼果的纵径和横径间具有显著性差异。根据PLS-DA判别模型累计训练集的分类正确率为99.59%,测试集的分类正确率为99.15%。说明主成分分析与光谱数据所建立的PLS-DA模型性能较好,对无花果成熟度的快速鉴别是有效且可行的。     

电子舌在中华绒螯蟹产地鉴别及等级评定的应用

采用电子舌对产自阳澄湖、松江、崇明的不同等级雌性中华绒螯蟹各可食部位的滋味轮廓进行检测。运用主成分分析法处理电子舌测定数据后发现,特级、1级、2级阳澄湖中华绒螯蟹体肉、钳肉、足肉、性腺4 个部位滋味轮廓区分显著。采用软独立建模的方法,分别建立了基于单部位及联合多部位的阳澄湖中华绒螯蟹产地鉴别模型,无论是单部位还是多部位模型,对非阳澄湖蟹样的拒绝率均为100%。采用偏最小二乘-判别分析法建立了中华绒螯蟹的等级评定模型,特级、1级、2级蟹样的偏最小二乘-判别分析模型相关系数均在0.90以上,采用上述模型可100%正确识别16 个未知蟹样的等级。     

Non-destructive prediction of hardening pericarp disorder in intact mangosteen by near infrared transmittance spectroscopy

A non-destructive technique to predict a hardening pericarp disorder in intact mangosteen is proposed by using near infrared (NIR) transmittance spectroscopy in the wavelength range of 660-960 nm. The study found that the spectral features of normal pericarp mangosteen and hardening pericarp mangosteen were different. The averaged spectra and individual spectra of hardening pericarp mangosteen from a calibration set (N = 560) were used to develop classification models, using partial least squares discriminant analysis (PLS-DA). A model based on individual spectra obtained better classification. The overall accuracy of classification for a prediction set (N = 358) was 91%. Out of 179 samples of normal pericarp fruits, 167 were identified correctly, while 159 samples out of 179 samples with hard pericarp were predicted correctly. The results showed that NIR transmittance spectroscopy can be used to predict hard pericarp disorder in intact mangosteen fruit accurately.     

Identification of Possible Milk Adulteration Using Physicochemical Data and Multivariate Analysis

Milk is one the most widely consumed foods in the world, with an average annual production of 723 million tons in the second half of the past decade. However, to increase milk’s profitability, some actors in the dairy chain may adulterate it, altering its chemical composition, and reducing the nutritional value of this food. The quality of milk is therefore assessed through chemical and physical analyses such as total dry matter, total Kjeldahl nitrogen, ash, acidity, fat, reducing sugars, depression of freezing point, and relative density. In this work, we have used Principal Components Analysis (PCA) and supervised methods (PLS-DA, SIMCA, kNN, and SVM-DA) to explore physicochemical data from milk and to classify and discriminate between samples that were compliant or not to the parameters set in the Brazilian Regulation for the Inspection of Animal Products and other national regulations. Classification results regarding specificity and selectivity for PLS-DA, SIMCA, kNN, and SVM-DA for noncompliant samples in the test group were, respectively, the following: 91, 88; 100, 97; 78, 67; 78, and 71%.     

Discrimination of Brazilian artisanal and inspected pork sausages: Application of unsupervised,linear and non-linear supervised chemometric methods

The performance of different chemometric approaches to discriminate artisanal and industrial pork sausages using traditional physicochemical parameters was investigated. A total of 90 samples of sausages marketed in various supermarkets and open-markets in Rio de Janeiro, Brazil were analyzed for their content of moisture, protein, fat, nitrite, sodium and calcium. Principal component analysis (PCA) and hierarchical cluster analysis (HCA) were used as exploratory methods, while linear and non-linear classification methods, such as k-nearest neighbors (k-NN), soft independent modeling of class analogy (SIMCA), partial least square discriminant analysis (PLSDA) and artificial neural networks (ANN) were used for assessing the data. Different behaviors for all parameters were analyzed between the classes. Principal component analysis and hierarchical cluster analysis did not show a complete discrimination of the samples. KNN and ANN results showed excellent performance for both categories with 100% correct prediction while SIMCA and PLSDA presented performance of 100% and 85.7% for inspected and artisanal sausages, respectively. According to the SIMCA, PLSDA and ANN, the contents of moisture and fat showed the highest discriminative power. Overall, the findings emphasize the use of multivariate techniques to evaluate the quality of processed foods, as pork sausages.     

Applications of FT-NIRS combined with PLS multivariate methods for the detection & quantification of saccharin adulteration in commercial fruit juices

Detection of adulteration in carbohydrate-rich foods like fruit juices is particularly difficult because of the variety of the commercial sweeteners available that match the concentration profiles of the major carbohydrates in the foods. In present study, a new sensitive and robust assay using Fourier Transform Near-Infrared Spectroscopy (FT-NIRS) combined with partial least square (PLS) multivariate methods has been developed for detection and quantification of saccharin adulteration in different commercial fruit juice samples. For this investigation, six different commercially available fruit juice samples were intentionally adulterated with saccharin at the following percentage levels: 0%, 0.10%, 0.30%, 0.50%, 0.70%, 0.90%, 1.10%, 1.30%, 1.50%, 1.70% and 2.00% (weight/volume). Altogether, 198 samples were used including 18 pure juice samples (unadulterated) and 180 juice samples adulterated with saccharin. PLS multivariate methods including partial least-squares discriminant analysis (PLS-DA) and partial least-squares regressions (PLSR) were applied to the obtained spectral data to build models. The PLS-DA model was employed to differentiate between pure fruit juice samples and those adulterated with saccharin. The R² value obtained for the PLS-DA model was 97.90% with an RMSE error of 0.67%. Similarly, a PLS regression model was also developed to quantify the amount of saccharin adulterant in juice samples. The R² value obtained for the PLSR model was 97.04% with RMSECV error of 0.88%. The employed model was then cross-validated by using a test set which included 30% of the total adulterated juice samples. The excellent performance of the model was proved by the low root mean squared error of prediction value of 0.92% and the high correlation factor of 0.97. This newly developed method is robust, nondestructive, highly sensitive and economical.     

Instrumental and multivariate statistical analyses for the characterisation of the geographical origin of Apulian virgin olive oils

In this paper, virgin olive oils (VOOs) coming from three different geographic origins of Apulia, were analysed for free acidity, peroxide value, spectrophotometric indexes, chlorophyll content, sterol, fatty acid, and triacylglycerol compositions. In order to predict the geographical origin of VOOs, different multivariate approaches were applied. By performing principal component analysis (PCA) a modest natural grouping of the VOOs was observed on the basis of their origin, and consequently three supervised techniques, i.e., general discriminant analysis (GDA), partial least squares-discriminant analysis (PLS-DA) and soft independent modelling of class analogy (SIMCA) were used and the results were compared. In particular, the best prediction ability was produced by applying GDA (average prediction ability of 82.5%), even if interesting results were obtained also by applying the other two classification techniques, i.e., 77.2% and 75.5% for PLS-DA and SIMCA, respectively.     

Application of mid-infrared spectroscopy with multivariate analysis and soft independent modeling of class analogies (SIMCA) for the detection of adulterants in minced beef

Chemometric MID-FTIR methods were developed to detect and quantify the adulteration of mince meat with horse meat, fat beef trimmings, and textured soy protein. Also, a SIMCA (Soft Independent Modeling Class Analogy) method was developed to discriminate between adulterated and unadulterated samples. Pure mince meat and adulterants (horse meat, fat beef trimmings and textured soy protein) were characterized based upon their protein, fat, water and ash content. In order to build the calibration models for each adulterant, mixtures of mince meat and adulterant were prepared in the range 2–90% (w/w). Chemometric analyses were obtained for each adulterant using multivariate analysis. A Partial Least Square (PLS) algorithm was tested to model each system (mince meat + adulterant) and the chemical composition of the mixture. The results showed that the infrared spectra of the samples were sensitive to their chemical composition. Good correlations between absorbance in the MID-FTIR and the percentage of adulteration were obtained in the region 1800–900 cm^− 1. Values of R² greater than 0.99, standard errors of calibration (SEC) in the range to 0.0001–1.278 and standard errors of prediction (SEP estimated) between 0.001 and 1.391 for the adulterant and chemical parameters were obtained. The SIMCA model showed 100% classification of adulterated meat samples from unadulterated ones. Chemometric MID-FTIR models represent an attractive option for meat quality screening without sample pretreatments which can identify the adulterant and quantify the percentage of adulteration and the chemical composition of the sample.