Class Modeling Techniques Research Articles

Classification of Natural and Thermo-Modified Peat by Class-Modeling Techniques

Significant attention is paid to the production of peat-based materials in modern days. The study explores the influence of natural peat thermal modification on its properties by applying class-modeling techniques. Modification of different types of peat is achieved by heating at 250 °C. The set of peat properties such as component composition, g-factor and IR-spectra is used to obtain data matrix. It is shown that class-modeling techniques, such as partial least-squares discriminant analysis (PLS-DA) and simple independent modeling of class analogy (SIMCA), allow estimating peat class (natural or modified) by a set of properties without prediction errors by using three latent variables. According to the results of classification, it is established that thermal modification can be considered as a means of regulating the composition and physico-chemical properties of natural peats as a raw material

Editorial

Many rural communities in Nepal experience considerable water stress during the dry season. Water storage systems (WSSs) have been proposed to supplement rain-fed irrigation and augment domestic water services in these communities. We evaluate household preferences for WSSs using a choice experiment and latent class modeling techniques. Results indicate the presence of three classes. The majority of households (≈92%) belong to two equally-sized classes, a relatively privileged group (i.e. wealthier, better educated, etc.) with strong preferences for supplemental irrigation and a less privileged group that is mainly interested in improved domestic water services. The remaining class’ preferences are dominated by the cost attribute and are consistent with households facing severe cash constraints. Estimated welfare effects reveal that WSSs disproportionately benefit the privileged, although this disparity is mitigated with the provision of domestic water. These findings highlight the potential welfare gains from WSS investments, but stress the need for multi-purpose water resource development and the potential for elite capture.

Water storage systems and preference heterogeneity in water-scarce environments: A choice experiment in Nepal’s Koshi River Basin

Abstract Many rural communities in Nepal experience considerable water stress during the dry season. Water storage systems (WSSs) have been proposed to supplement rain-fed irrigation and augment domestic water services in these communities. We evaluate household preferences for WSSs using a choice experiment and latent class modeling techniques. Results indicate the presence of three classes. The majority of households (≈92%) belong to two equally-sized classes, a relatively privileged group (i.e. wealthier, better educated, etc.) with strong preferences for supplemental irrigation and a less privileged group that is mainly interested in improved domestic water services. The remaining class’ preferences are dominated by the cost attribute and are consistent with households facing severe cash constraints. Estimated welfare effects reveal that WSSs disproportionately benefit the privileged, although this disparity is mitigated with the provision of domestic water. These findings highlight the potential welfare gains from WSS investments, but stress the need for multi-purpose water resource development and the potential for elite capture.

Application of class-modelling techniques to infrared spectra for analysis of pork adulteration in beef jerkys.

The use of chemometrics to analyse infrared spectra to predict pork adulteration in the beef jerky (dendeng) was explored. In the first step, the analysis of pork in the beef jerky formulation was conducted by blending the beef jerky with pork at 5-80% levels. Then, they were powdered and classified into training set and test set. The second step, the spectra of the two sets was recorded by Fourier Transform Infrared (FTIR) spectroscopy using atenuated total reflection (ATR) cell on the basis of spectral data at frequency region 4000-700cm(-1). The spectra was categorised into four data sets, i.e. (a) spectra in the whole region as data set 1; (b) spectra in the fingerprint region (1500-600cm(-1)) as data set 2; (c) spectra in the whole region with treatment as data set 3; and (d) spectra in the fingerprint region with treatment as data set 4. The third step, the chemometric analysis were employed using three class-modelling techniques (i.e. LDA, SIMCA, and SVM) toward the data sets. Finally, the best result of the models towards the data sets on the adulteration analysis of the samples were selected and the best model was compared with the ELISA method. From the chemometric results, the LDA model on the data set 1 was found to be the best model, since it could classify and predict 100% accuracy of the sample tested. The LDA model was applied toward the real samples of the beef jerky marketed in Jember, and the results showed that the LDA model developed was in good agreement with the ELISA method.

Discriminating the Geographical Origins of Chinese White Lotus Seeds by Near-Infrared Spectroscopy and Chemometrics

The traceability of a Chinese white lotus seed (WLS) with Protected Designation of Origin (PDO) was investigated using near-infrared (NIR) spectroscopy and chemometrics. Three chemometrics methods, discrimination analysis (DA), class modeling, and a newly proposed strategy, the fusion of DA and class modeling, were investigated to compare their capacity to trace the geographical origins of WLS. Least squares support vector machine (LS-SVM) was developed to distinguish the PDO WLS from non-PDO WLS of four main producing areas. A class modeling technique, one-class partial least squares (OCPLS), was developed only using the data of PDO WLS. By the fusion of LS-SVM and OCPLS, the best prediction sensitivity and specificity were 0.900 and 0.973, respectively. The results indicate that fusion of DA and class modeling can enhance the specificity for detection of non-PDO products. The conclusion is that DA and class modeling should be combined for tracing food geographical origins.

Poly-traumatization and harmful behaviors in a sample of emergency department Psychiatric Intake Response Center youth

The purpose of this study is to examine risk factors for poly-traumatization, and the impact of poly-traumatization on harmful behaviors (suicidal, self-harm, and violent), among a group of pediatric patients presenting at an emergency department's psychiatric intake response center. We employed a retrospective medical chart review in a children's hospital for a 2-year span (N=260). The study employed 2 statistical analyses. The first analysis used multinomial logistic regression to model the odds of harmful behaviors comparing increasing numbers of co-occurring traumatization types. The second analysis employed latent class modeling techniques in three ways to (a) define different poly-traumatization populations, (b) examine the relationship between predictors and class assignment, and (c) examine the relationship between class assignment and harmful behavioral outcomes. About 62% of the sample presented with at least 1 traumatization type and about 50% one harmful behavior type. Compared to those with 1, 2, or 3 traumatization types, patients with 4 or more traumatization types have higher odds of harmful behaviors. The latent class analysis revealed 2 populations: High serious victimization and minimal traumatization. History of family mental health issues was the only significant predictor of class membership. Class membership was associated with all of the harmful behavioral outcome categories. These findings support consideration of poly-traumatization as a risk factor for the high occurrence of harmful behaviors in this sample of pediatric psychiatric patients and that history of family mental health issues may contribute to the high co-occurrence of poly-traumatization.

Estimating Benchmark Exposure for Air Particulate Matter Using Latent Class Models

We performed benchmark exposure (BME) calculations for particulate matter when multiple dichotomous outcome variables are involved using latent class modeling techniques and generated separate results for both the extra risk and additional risk. The use of latent class models in this study is advantageous because it combined several outcomes into just two classes (namely, a high-risk class and a low-risk class) and compared these two classes to obtain the BME levels. This novel approach addresses a key problem in risk estimation--namely, the multiple comparisons problem, where separate regression models are fitted for each outcome variable and the reference exposure will rely on the results of the best-fitting model. Because of the complex nature of the estimation process, the bootstrap approach was used to estimate the reference exposure level, thereby reducing uncertainty in the obtained values. The methodology developed in this article was applied to environmental data by identifying unmeasured class membership (e.g., morbidity vs. no morbidity class) among infants in utero using observed characteristics that included low birth weight, preterm birth, and small for gestational age.

Class-modeling approach to PTR-TOFMS data: a peppers case study.

Proton transfer reaction-mass spectrometry (PTR-MS), in its recently developed implementation based on time-of-flight mass spectrometry (PTR-TOFMS), was used to rapidly determine the volatile compounds present in fruits of Capsicum spp. We analyzed the volatile organic compounds emission profile of freshly cut chili peppers belonging to three species and 33 different cultivars. PTR-TOFMS data, analyzed with appropriate and advanced multivariate class-modeling approaches, perfectly discriminated among the three species (100% correct classification in validation set). VIP (variable importance in projection) scores were used to select the 15 most important volatile compounds in discriminating the species. The best candidates for Capsicum spp. were compounds with measured m/z of 63.027, 101.096 and 107.050, which were, respectively, tentatively identified as dimethyl sulfide, hexanal and benzaldehyde. Based on the promising results, the possibility of introducing multivariate class-modeling techniques, different from the classification approaches, in the field of volatile compounds analyses is discussed.

Rapid analysis of adulterations in Chinese lotus root powder (LRP) by near-infrared (NIR) spectroscopy coupled with chemometric class modeling techniques

This paper develops a rapid analysis method for adulteration identification of a popular traditional Chinese food, lotus root powder (LRP), by near-infrared spectroscopy and chemometrics. 85 pure LRP samples were collected from 7 main lotus producing areas of China to include most if not all of the significant variations likely to be encountered in unknown authentic materials. To evaluate the model specificity, 80 adulterated LRP samples prepared by blending pure LRP with different levels of four cheaper and commonly used starches were measured and predicted. For multivariate quality models, two class modeling methods, the traditional soft independent modeling of class analogy (SIMCA) and a recently proposed partial least squares class model (PLSCM) were used. Different data preprocessing techniques, including smoothing, taking derivative and standard normal variate (SNV) transformation were used to improve the classification performance.The results indicate that smoothing, taking second-order derivatives and SNV can improve the class models by enhancing signal-to-noise ratio, reducing baseline and background shifts. The most accurate and stable models were obtained with SNV spectra for both SIMCA (sensitivity 0.909 and specificity 0.938) and PLSCM (sensitivity 0.909 and specificity 0.925). Moreover, both SIMCA and PLSCM could detect LRP samples mixed with 5% (w/w) or more other cheaper starches, including cassava, sweet potato, potato and maize starches. Although it is difficult to perform an exhaustive collection of all pure LRP samples and possible adulterations, NIR spectrometry combined with class modeling techniques provides a reliable and effective method to detect most of the current LRP adulterations in Chinese market.

Valuing improvements to coastal waters using choice experiments: An application to revisions of the EU Bathing Waters Directive

Planned changes to the European Union’s Bathing Waters Directive (2006/7EC) will force member states to produce improvements in a number of parameters of coastal water quality. This study uses the choice experiment method to estimate the economic benefits attached to such improvements, based on a sample of recreationalists on beaches in Ireland. The analysis indicates that improvements in all of the bathing water related attributes studied result in positive willingness to pay, and also show evidence of scope effects. Using random parameters and latent class modelling techniques, potential heterogeneity in preferences is then investigated and shown to be present to a significant degree. One observable determinant of this preference heterogeneity is the degree of exposure of individuals to health risks relating to water quality, as proxied by the type of recreational activity they undertake.

The Feasibility of Using Near-Infrared Spectroscopy and Chemometrics for Untargeted Detection of Protein Adulteration in Yogurt: Removing Unwanted Variations in Pure Yogurt

Untargeted detection of protein adulteration in Chinese yogurt was performed using near-infrared (NIR) spectroscopy and chemometrics class modelling techniques. sixty yogurt samples were prepared with pure and fresh milk from local market, and 197 adulterated yogurt samples were prepared by blending the pure yogurt objects with different levels of edible gelatin, industrial gelatin, and soy protein powder, which have been frequently used for yogurt adulteration. A recently proposed one-class partial least squares (OCPLS) model was used to model the NIR spectra of pure yogurt objects and analyze those of future objects. To improve the raw spectra, orthogonal projection (OP) of raw spectra onto the spectrum of pure water and standard normal variate (SNV) transformation were used to remove unwanted spectral variations. The best model was obtained with OP preprocessing with sensitivity of 0.900 and specificity of 0.949. Moreover, adulterations of yogurt with 1% (w/w) edible gelatin, 2% (w/w) industrial gelatin, and 2% (w/w) soy protein powder can be safely detected by the proposed method. This study demonstrates the potential of combining NIR spectroscopy and OCPLS as an untargeted detection tool for protein adulteration in yogurt.

Protected Geographical Indication Identification of a Chinese Green Tea (Anji-White) by Near-Infrared Spectroscopy and Chemometric Class Modeling Techniques

This paper reports a rapid identification method for a Chinese green tea with PGI, Anji-white tea, by class modeling techniques and NIR spectroscopy. 167 real and representative Anji-white tea samples were collected from 8 tea plantations in their original producing areas for model training. Another 81 non-Anji-white tea samples of similar appearance were collected from 7 important tea producing areas and used for validation of model specificity. Diffuse NIR spectra were measured with finely ground tea powders. OCPLS and SIMCA were used to describe the distribution of representative Anji-white tea objects and predict the authenticity of new objects. For data preprocessing, smoothing, derivatives, and SNV were applied to improve the raw spectra and classification performance. It is demonstrated that taking derivatives and SNV can improve classification accuracy and reduce the complexity of class models by removing spectral background and baseline. For the best models, the sensitivity and specificity were 0.886 and 0.951 for OCPLS, 0.886 and 0.938 for SIMCA with SNV spectra, respectively. Although it is difficult to perform an exhaustive analysis of all types of potential false objects, the proposed method can detect most of the important non-Anji-white teas in the Chinese market.

Combining local wavelength information and ensemble learning to enhance the specificity of class modeling techniques: Identification of food geographical origins and adulteration

Class modeling techniques are required to tackle various one-class problems. Because the training of class models is based on the target class and the origins of future test objects usually cannot be exactly predefined, the criteria for feature selection of class models are not very straightforward. Although feature reduction can be expected to improve class models performance, more features retained can provide a sufficient description of the sought-for class. This paper suggests a strategy to balance class description and model specificity by ensemble learning of sub-models based on separate local wavelength intervals. The acceptance or rejection of a future object can be explicitly determined by examining its acceptance frequency by sub-models. Considering the lack of information about sub-model independence, we propose to use a data-driven method to control the sensitivity of the ensemble model by cross validation. In this way, all the wavelength intervals are used for class description and the local wavelength intervals are highlighted to enhance the ability to detect out-of-class objects.The proposed strategy was performed on one-class partial least squares (OCPLS) and soft independent modeling of class analogy (SIMCA). By analysis of two infrared spectral data sets, one for geographical origin identification of white tea and the other for discrimination of adulterations in pure sesame oil, the proposed ensemble class modeling method was demonstrated to have similar sensitivity and better specificity compared with total-spectrum SIMCA and OCPLS models. The results indicate local spectral information can be extracted to enhance class model specificity.

Rapid geographical origin analysis of pure West Lake lotus root powder (WL-LRP) by near-infrared spectroscopy combined with multivariate class modeling techniques

A rapid analysis method was proposed for the identification of a popular traditional Chinese food with Protected Geographical Indication (PGI), West Lake lotus root powder (WL-LRP), by near-infrared spectroscopy and chemometric methods. A set of 105 pure and real WL-LRP samples were collected from 9 main producers to obtain a representative training set of the authentic objects. 95 non-WL-LRP samples from 8 different main lotus producing areas of China were analyzed for validation of model specificity. Linear partial least squares class model (PLSCM) and nonlinear support vector data description (SVDD) were used to develop quality control models of authentic and pure WL-LRP objects. Spectral data were preprocessed by taking derivatives and standard normal variate (SNV) transformation. The analysis results indicate performing SNV transformation can obtain more stable and accurate models compared with taking derivatives. The best models were obtained with SNV spectra, the sensitivity and specificity was 0.87 and 0.90 for PLSCM, 0.93 and 0.92 for SVDD, respectively.

FTIR Spectroscopy and Chemometric Class Modeling Techniques for Authentication of Chinese Sesame Oil

AbstractThis investigation was aimed at developing a rapid analysis method for authentication of Chinese sesame oils by FTIR spectrometry and chemometrics. Ninety‐five sesame oil samples were collected from the six main producing areas of China to include most if not all of the significant spectral variations likely to be encountered in future authentic materials. Two class modeling techniques, the soft independent modeling of class analogy (SIMCA) and the partial least squares class model (PLSCM) were investigated and the data preprocessing techniques including smoothing, derivative and standard normal variate (SNV) tests were performed to improve the classification performance. It was demonstrated that SIMCA and PLSCM can detect various adulterated sesame oils doped with 3% or more (w/w) of other cheaper oils, including rapeseed, soybean, palm and peanut oils. First derivative, second derivative and SNV tests significantly enhanced the class models by reducing baseline and background shifts. Smoothing of raw spectra led to inferior identification performance and proved itself to be unsuitable because some of the detailed frequency details were lost during smoothing. The best model performance was obtained with second derivative spectra by SIMCA (sensitivity 0.905 and specificity 0.944) and PLSCM (sensitivity 0.952 and specificity 0.937). Although it is difficult to perform an exhaustive sampling of all types of pure sesame oils and potential adulterations, PLS and SIMCA combined with FTIR spectrometry can detect most of current adulterations of sesame oils on the Chinese market.

Class Modeling Analysis of Heparin 1H NMR Spectral Data Using the Soft Independent Modeling of Class Analogy and Unequal Class Modeling Techniques

To differentiate heparin samples with varying amounts of dermatan sulfate (DS) impurities and oversulfated chondroitin sulfate (OSCS) contaminants, proton NMR spectral data for heparin sodium active pharmaceutical ingredient samples from different manufacturers were analyzed using multivariate chemometric techniques. A total of 168 samples were divided into three groups: (a) Heparin, [DS] ≤ 1.0% and [OSCS] = 0%; (b) DS, [DS] > 1.0% and [OSCS] = 0%; (c) OSCS, [OSCS] > 0% with any content of DS. The chemometric models were constructed and validated using two well-established methods: soft independent modeling of class analogy (SIMCA) and unequal class modeling (UNEQ). While SIMCA modeling was conducted using the entire set of variables extracted from the NMR spectral data, UNEQ modeling was combined with variable reduction using stepwise linear discriminant analysis to comply with the requirement that the number of samples per class exceed the number of variables in the model by at least 3-fold. Comparison of the results from these two modeling approaches revealed that UNEQ had greater sensitivity (fewer false positives) while SIMCA had greater specificity (fewer false negatives). For Heparin, DS, and OSCS, respectively, the sensitivity was 78% (56/72), 74% (37/50), and 85% (39/46) from SIMCA modeling and 88% (63/72), 90% (45/50), and 91% (42/46) from UNEQ modeling. Importantly, the specificity of both the SIMCA and UNEQ models was 100% (46/46) for Heparin with respect to OSCS; no OSCS-containing sample was misclassified as Heparin. The specificity of the SIMCA model (45/50, or 90%) was superior to that of the UNEQ model (27/50, or 54%) for Heparin with respect to DS samples. However, the overall prediction ability of the UNEQ model (85%) was notably better than that of the SIMCA model (76%) for the Heparin vs DS vs OSCS classes. The models were challenged with blends of heparin spiked with nonsulfated, partially sulfated, or fully oversulfated chondroitin sulfate A, dermatan sulfate, or heparan sulfate at the 1.0, 5.0, and 10.0 wt % levels. The results from the present study indicate that the combination of (1)H NMR spectral data and class modeling techniques (viz., SIMCA and UNEQ) represents a promising strategy for assessing the quality of commercial heparin samples with respect to impurities and contaminants. The methodologies show utility for applications beyond heparin to other complex products.

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%.

Comparison between classical and innovative class-modelling techniques for the characterisation of a PDO olive oil

An authentication study of the Italian PDO (protected designation of origin) olive oil Chianti Classico, based on near-infrared and UV-Visible spectroscopy, an artificial nose and an artificial tongue, with a set of samples representative of the whole Chianti Classico production and a considerable number of samples from a close production area (Maremma) was performed. The non-specific signals provided by the four fingerprinting analytical techniques, after a proper pre-processing, were used for building class models for Chianti Classico oils. The outcomes of classical class-modelling techniques like soft independent modelling of class analogy and quadratic discriminant analysis-unequal dispersed classes were compared with those of two techniques recently introduced into Chemometrics: multivariate range modelling and CAIMAN analogues modelling methods.

Two class-modelling techniques that give families of class-models and their relation with the structure of the data

In class-modelling problems, which are again becoming increasingly important, there are two parameters to value the quality of the class-model built for a category, namely sensitivity and specificity. Using them as criteria, in this paper, two different approaches to class-modelling problems are presented, approaches that differ from other usual methods in the fact that they provide not just one class-model per category but a set of different class-models that accounts for the possible pairs of sensitivity-specificity values attainable for a given data set. One of the proposals is partial least squares class-modelling (PLS-CM) that, by the joint use of PLS with binary responses and the posterior statistical modelling of the distribution of the computed responses, permits the estimation of the risks related to the decision of assigning a sample into a class, and thus, the values of sensitivity and specificity. The other proposed method, Pareto-optimal front in class-modelling, is an analytical approach posed in a multi-response optimization framework, the one that corresponds to trying to simultaneously maximise the sensitivity and specificity of a class-model. Additionally, the whole family of computed class-models is validated in prediction by using cross-validation, showing the stability of both methods for prediction. The case-studies show the complementariness of both approaches and, in particular, that the joint use of both techniques allows the user to detect possible structures in the data set especially inadequate for PLS. The results, i.e. the whole set of sensitivity-specificity values achievable for a given problem, are graphically represented to improve its study and make it easy to make a decision about the model.

Application of class-modelling techniques to near infrared data for food authentication purposes

Following the introduction of legal identifiers of geographic origin within Europe, methods for confirming any such claims are required. Spectroscopic techniques provide a method for rapid and non-destructive data collection and a variety of chemometric approaches have been deployed for their interrogation. In this present study, class-modelling techniques (SIMCA, UNEQ and POTFUN) have been deployed after data compression by principal component analysis for the development of class-models for a set of olive oils and honeys. The number of principal components, the confidence level and spectral pre-treatments (1st and 2nd derivative, standard normal variate) were varied, and a strategy for variable selection was tried. Models were evaluated on a separate validation sample set. The outcomes are reported and criteria for selection of the most appropriate models for any given application are discussed.