Recognition of organic rice samples based on trace elements and support vector machines

Elemental profile of wheat in the las vegas market: Geographic origin discrimination and probabilistic health risk assessment

This study investigates concentrations of toxic and potentially toxic elements (PTEs) in organic and conventional wheat flour and grains marketed in Las Vegas. Geographic origins of the samples were evaluated using Linear Discriminant Analysis (LDA). Monte Carlo Simulation technique was also employed to evaluate non-carcinogenic risk in four life stages. Concentrations of Al, As, Cd, Co, Cr, Cu, Fe, Mn, Mo, Ni, Pb, Se, Sr, and Zn were determined using inductively coupled plasma mass spectrometry (ICP-MS) following hot block-assisted digestion. Obtained results showed non-significant differences in contents of toxic and PTEs between conventional and organic wheat grains/flour. Using LDA, metal (loid)s were found to be indicative of geographical origin. The LDA produced a total correct classification rate of 95.8% and 100% for US and West Pacific Region samples, respectively. The results of the present study indicate that the estimated non-carcinogenic risk associated with toxic element intakes across the four life stages were far lower than the threshold value (Target Hazard Quotient (THQ) > 1). However, the probability of exceeding the threshold value for Mn is approximately 32% in children aged between 5 and 8 years. The findings of this study can aid in understanding dietary Mn exposure in children in Las Vegas.

Metal(loid)s Profile of Four Traditional Ethiopian Teff Brands: Geographic Origin Discrimination

Among the most renowned Ethiopian food crops, teff (Eragrostis tef (Zucc.)Trotter) is the most nutritious and gluten-free cereal. Because of the increase in demand for teff, it is necessary to establish geographic origin authentication of traditional teff brands based on multi-element fingerprint. For this purpose, a total of 60 teff samples were analysed using Inductively Coupled Plasma Mass Spectrometry (ICP-MS). Accuracy of the laboratory procedure was verified by the analysis of rice flour standard reference material (NIST SRM 1568b). In this context, four traditional teff brands (Ada'a, Ginchi, Gojam and Tulu Bolo) were analytically characterized using multi-element fingerprint and further treated statistically using linear discriminant analysis (LDA). Due to obvious extrinsic Fe, Al and V contamination, these elements were excluded from the discriminant model. Five elements (Cu, Mo, Se, Sr, and Zn) significantly contributed to discriminate the geographical origin of white teff. On the other hand, Mn, Mo, Se and Sr were used as discriminant variables for brown teff. LDA revealed 90 and 100% correct classifications for white and brown teff, respectively. Overall, multi-element fingerprint coupled with LDA can be considered a suitable tool for geographic origin discrimination of traditional teff brands.

A comparative study regarding the adulteration detection of honey: Physicochemical parameters vs. impedimetric data

Honey adulteration is a major issue for European Union and its members because of an unfair practice of different producers and beekeepers, many adulterations involve the addition of sweet, concentrated syrups which may appear like honey. In our study we analysed the influence of adulteration of tilia honey with different syrups (such as corn, rice, inverted sugar, agave, maple syrups) in different percentages (5%, 10%, and 20% respectively) on physicochemical parameters (moisture content, L*, hab,cab, pH, free acidity, electrical conductivity (EC), 5-hydroxymetilfurfural (HMF), fructose, glucose, sucrose, turanose, trehalose, melesitose and raffinose) and impedimetric properties using electrochemical impedance spectroscopy. The impedimetric sensing was made using an electrochemical cell composed of two gold electrodes, and the frequency ranged between 0.1 kHz and 100 kHz. The impedimetric parameters (Z', Z″ and phase) and Randal circuit parameters can distinguish the authentic honeys from the adulterated ones (based on the adulteration agent and adulteration percentage, respectively). The partial least squares - discriminant analysis (PLS-DA) and support vector machines (SVM) were used in a binary mode to separate the authentic honeys from the adulterated ones, and the SVM proved to separate much better than PLS-DA.

Use of elemental profiles determined by energy-dispersive X-ray fluorescence and multivariate analyses to detect adulteration in Ceylon cinnamon

The price of Cinnamomum verum (Ceylon cinnamon) is around twice as high as that of the other cinnamon varieties commonly grouped under the name cassia cinnamon, making the former spice an attractive target for fraudsters. This work demonstrates that elemental profiles obtained by energy-dispersive X-ray fluorescence in combination with multivariate analyses can be used as a screening method to detect Ceylon cinnamon adulteration. Thirty-six elements were analysed in 52 commercially available cinnamon samples, 29 Ceylon, 8 cassia, and 15 for which no indication about variety was provided. Fifty-eight percent of the samples were either adulterated or did not meet international quality criteria. Four of the ground cinnamon samples labelled as Ceylon cinnamon were found to be pure cassia or a mixture with a high cassia content, and 26 samples were suspected of other types of adulteration including replacement of bark with other parts of the cinnamon tree. Headspace gas chromatography-mass spectrometry and ash determination by thermogravimetric analysis confirmed the conclusions reached by elemental analysis. Only one sample labelled as Ceylon cinnamon and that according to its volatile composition was cassia cinnamon was not flagged as suspicious by elemental analysis.

Honey authentication and adulteration detection using emission - excitation spectra combined with chemometrics

The aim of this study was to evaluate the usefulness of emission-excitation matrices for honey authentication and adulteration detection. For this purpose, 4 types of authentic honeys (tilia, sunflower, acacia and rape) and samples adulterated with different adulteration agents (agave, maple, inverted sugar, corn and rice in different percentages - 5%, 10% and 20%) were analysed. Each honey type and each adulteration agent exhibit unique emission-excitation spectra that can be used for the classification according to the botanical origin and for the detection of adulteration. The principal component analysis clearly separated the rape, sunflower and acacia honeys. The partial least squares - discriminant analysis (PLS-DA) and support vector machines (SVM) were used in a binary mode to separate the authentic honeys from the adulterated ones, and the SVM proved to separate much better than PLS-DA.

Organic Rice Production Practices: Effects on Grain End-Use Quality, Healthfulness, and Safety

Demand for rice labeled as organic is growing globally. Consumers state that foods labeled as organic are nutritionally superior and safer than their conventionally produced equivalent. The research question for this systematic review is as follows: is there a difference between the effects of sustainable agriculture and conventional farming methods on rice grain end-use quality, healthfulness, and safety? The studies (n = 23) examined for this review suggest that organic production practices don't influence most end-use quality (e.g., chalk, milling yield, pasting properties) and healthfulness (e.g., gamma-oryzanol fraction and tocols) traits studied, or if it does, it will be to a small degree. If differences in end-use quality traits are found, they will be associated with grain protein content, which varies along with the dose of nitrogen applied during rice growth. We conclude that the studies evaluated in this review found that organically produced rice grain was less likely to contain residues of the pesticides (e.g., organochlorine) examined in the study than the rice is grown using conventional methods. There was some evidence that organically grown rice is more likely to be contaminated with mycotoxin-producing fungi and some mycotoxins. Common shortcomings of the studies were that they were poorly designed, with limited to no details of the cultural management practices used to grow the rice studied, the length of time fields was under organic management not stated, cultivars were not named, and the data wasn't analyzed statistically.

Deep Learning Combined with Hyperspectral Imaging Technology for Variety Discrimination of Fritillaria thunbergii

Traditional Chinese herbal medicine (TCHM) plays an essential role in the international pharmaceutical industry due to its rich resources and unique curative properties. The flowers, stems, and leaves of Fritillaria contain a wide range of phytochemical compounds, including flavonoids, essential oils, saponins, and alkaloids, which may be useful for medicinal purposes. Fritillaria thunbergii Miq. Bulbs are commonly used in traditional Chinese medicine as expectorants and antitussives. In this paper, a feasibility study is presented that examines the use of hyperspectral imaging integrated with convolutional neural networks (CNN) to distinguish twelve (12) Fritillaria varieties (n = 360). The performance of support vector machines (SVM) and partial least squares-discriminant analysis (PLS-DA) was compared with that of convolutional neural network (CNN). Principal component analysis (PCA) was used to assess the presence of cluster trends in the spectral data. To optimize the performance of the models, cross-validation was used. Among all the discriminant models, CNN was the most accurate with 98.88%, 88.89% in training and test sets, followed by PLS-DA and SVM with 92.59%, 81.94% and 99.65%, 79.17%, respectively. The results obtained in the present study revealed that application of HSI in conjunction with the deep learning technique can be used for classification of Fritillaria thunbergii varieties rapidly and non-destructively.

Geographic Origin Discrimination of Millet Using Vis-NIR Spectroscopy Combined with Machine Learning Techniques

Millet is a primary food for people living in the dry and semi-dry regions and is dispersed within most parts of Europe, Africa, and Asian countries. As part of the European Union (EU) efforts to establish food originality, there is a global need to create Protected Geographical Indication (PGI) and Protected Designation of Origin (PDO) of crops and agricultural products to ensure the integrity of the food supply. In the present work, Visible and Near-Infrared Spectroscopy (Vis-NIR) combined with machine learning techniques was used to discriminate 16 millet varieties (n = 480) originating from various regions of China. Five different machine learning algorithms, namely, K-nearest neighbor (K-NN), Linear discriminant analysis (LDA), Logistic regression (LR), Random Forest (RF), and Support vector machine (SVM), were used to train the NIR spectra of these millet samples and to assess their discrimination performance. Visible cluster trends were obtained from the Principal Component Analysis (PCA) of the spectral data. Cross-validation was used to optimize the performance of the models. Overall, the F-Score values were as follows: SVM with 99.5%, accompanied by RF with 99.5%, LDA with 99.5%, K-NN with 99.1%, and LR with 98.8%. Both the linear and non-linear algorithms yielded positive results, but the non-linear models appear slightly better. The study revealed that applying Vis-NIR spectroscopy assisted by machine learning technique can be an essential tool for tracing the origins of millet, contributing to a safe authentication method in a quick, relatively cheap, and non-destructive way.

Tracing the Geographical Origin of Thai Hom Mali Rice in Three Contiguous Provinces of Thailand Using Stable Isotopic and Elemental Markers Combined with Multivariate Analysis

Rice is a staple food for more than half of the world’s population. The discrimination of geographical origin of rice has emerged as an important issue to prevent mislabeling and adulteration problems and ensure food quality. Here, the discrimination of Thai Hom Mali rice (THMR), registered as a European Protected Geographical Indication (PGI), was demonstrated. Elemental compositions (Mn, Rb, Co, and Mo) and stable isotope (δ¹⁸O) in the rice were analyzed using inductively coupled plasma mass spectrometry (ICP-MS) and elemental analyzer isotope ratio mass spectrometry (EA-IRMS), respectively. The recoveries and precisions of all elements were greater than 98% and lower than 9%, respectively. The analytical precision (±standard deviation) was below ±0.2‰ for δ¹⁸O measurement. Mean of Mn, Rb, Co, Mo, and δ¹⁸O levels was 14.0 mg kg⁻¹, 5.39 mg kg⁻¹, 0.049 mg kg⁻¹, 0.47 mg kg⁻¹, and 25.22‰, respectively. Only five valuable markers combined with radar plots and multivariate analysis, linear discriminant analysis (LDA) could distinguish THMR cultivated from three contiguous provinces with correct classification and cross-validation of 96.4% and 92.9%, respectively. These results offer valuable insight for the sustainable management and regulation of improper labeling regarding geographical origin of rice in Thailand and other countries.

Voltammetric E-Tongue for Honey Adulteration Detection

The aim of this study is to establish the usefulness of an electronic tongue based on cyclic voltammetry e-tongue using five working electrodes (gold, silver, copper, platinum and glass) in honey adulteration detection. Authentic honey samples of different botanical origin (acacia, tilia, sunflower, polyfloral and raspberry) were adulterated with agave, maple, inverted sugar, corn and rice syrups in percentages of 5%, 10%, 20% and 50%. The silver and copper electrodes provided the clearest voltammograms, the differences between authentic and adulterated honey samples being highlighted by the maximum current intensity. The electronic tongue results have been correlated with physicochemical parameters (pH, free acidity, hydroxymethylfurfural content—5 HMF and electrical conductivity—EC). Using statistical methods such as Linear discriminant analysis (LDA) and Support vector machines (SVM), an accuracy of 94.87% and 100% respectively was obtained in the calibration step and 89.65% and 100% respectively in the validation step. The PLS-R (Partial Least Squares Regression) model (constructed from the minimum and maximum current intensity obtained for all electrodes) was used in physicochemical parameters prediction; EC reached the highest regression coefficients (0.840 in the calibration step and 0.842 in the validation step, respectively), being followed by pH (0.704 in the calibration step and 0.516 in the validation step, respectively).

Development of an Ozone-Assisted Sample Preparation Method for the Determination of Cu and Zn in Rice Samples

A green analytical method for the determination of Cu and Zn in rice samples was developed. This method was based on an ozone-assisted extraction (OAE) in diluted acid media. A novel closed system was designed for this purpose that allowed four simultaneous sample treatments being safe for the laboratory environment. The method consisted in 0.5 g of the sample, 15 minutes of ozonation, and 3 minutes of centrifugation. The obtained supernatant was ready for Cu and Zn determinations by flame atomic absorption spectrometry. Detection limits were 0.20 and 0.08 mg kg^-1 for Cu and Zn, respectively, with a precision (RSD) better than 5% for both elements. A certified reference material of rice flour was analyzed for trueness evaluation, and the mean recoveries (%) were 100.4 (Cu) and 95.9 (Zn). Several commercial rice samples were analyzed using this method, and the results were compared with those obtained using traditional microwave-assisted digestion (MAE). Both methods yielded comparable results. Cu and Zn levels were in accordance with reported values in other regions. The OAE resulted to be simple and economical and with results equivalent with those obtained using traditional sample preparation procedures as MAE with the advantage of being in good agreement with the principles of green analytical chemistry.

Feasibility study on the use of elemental profiles to authenticate aromatic rice: the case of Basmati and Thai rice

Among the thousands of existing rice varieties, aromatic rice has increasingly attracted consumer’s preference in recent years. Within aromatic rice, Basmati, cultivated in some regions in Pakistan and India, is highly demanded. Other aromatic rice, cultivated in specific regions, for instance in Thailand (commonly referred to as Jasmine Thai rice), are also highly appreciated by consumers. In this work, the elemental profiles of commercially available rice samples (17 Basmati, 11 Thai, and 7 Long Grain rice) were determined by energy-dispersive X-ray fluorescence (ED-XRF) spectroscopy. The mass fractions of P, Cl, S, K, Fe, Cu, and Zn were significantly different (95% confidence interval) between Basmati and Thai rice and between Thai and Long Grain rice; only Cl, S, and Zn were significantly different between Basmati and Long Grain rice. Multivariate evaluation of the results combining soft independent modelling by class analogy (SIMCA) and partial least square discriminant analysis (PLS-DA) allowed the correct classification (true positives) of 94.1, 85.6, and 100% of the Basmati, Long Grain, and Thai rice, respectively. The specificity (true negatives) of Basmati, Long Grain, and Thai was 94.4, 82.1, and 100%, respectively.

Are the elemental fingerprints of organic and conventional food different? ED-XRF as screening technique

Research has been conducted the last years to assess whether organically grown food is chemically different from produce of conventional agriculture and which markers are appropriate to discriminate between them. Most articles focus on one single food commodity, produced under strict controlled organic farming conditions, leaving open the question whether the difference would be seen when applied to the same commodity under different growing conditions. In this work 118 organic and 151 conventional samples of commercially available paprika powder, cinnamon, coffee, tea, chocolate, rice, wheat flour, cane sugar, coconut water, honey and bovine milk were characterised for their elemental composition using energy dispersive X-ray fluorescence. Resulting profiles were analysed using univariate and multivariate statistical techniques. Organic samples of a given commodity clustered together and were separated from their conventional counterparts. Differences in the elemental composition of food, could be used to develop statistical models for verifying the agronomical production system.

Authentication of PDO paprika powder (Pimentón de la Vera) by multivariate analysis of the elemental fingerprint determined by ED-XRF. A feasibility study

Products with a Protected Denomination of Origin (PDO) are vulnerable to misdescription of their true geographical origin. In this work a method has been developed that allows the authentication of La Vera paprika powder (Pimentón de la Vera), a PDO product from the central-west Spanish region, Extremadura. The mass fractions of Br, Ca, Cr, Cl, Cu, Fe, K, Mn, Ni, P, Rb, S, Sr and Zn determined by energy dispersive X-ray fluorescence (ED-XRF) are used for classification purposes by multivariate analysis using Soft Independent Modelling of Class Analogy (SIMCA) (PCA-Class) and Partial Least Square-Discriminant Analysis (PLS-DA). Sixty-seven paprika samples purchased in supermarkets around Europe and on-line via the official web-site of Pimentón de La Vera, were used to build up the models for prediction purposes. The PCA-class model of La Vera paprika powder had a sensitivity of 82%, a specificity of 100% and an accuracy of 91%, whereas the PLS-DA model had a sensitivity of 100%, a specificity of 91% and an accuracy of 96%.

•

Authentication of Pimentón de la Vera is achieved by ED-XRF elemental fingerprint.

•

ED-XRF is fast and hardly requires any sample treatment.

•

Hazardous reagents are not required and chemical waste is not generated.

•

SIMCA and PLS-DA models are fit-for-the purpose of fighting food fraud.

Authentication of Rice (Oryza sativa L.) Using Near Infrared Spectroscopy Combined with Different Chemometric Classification Strategies

Rice is a staple food in Vietnam, and the concern about rice is much greater than that for other foods. Preventing fraud against this product has become increasingly important in order to protect producers and consumers from possible economic losses. The possible adulteration of this product is done by mixing, or even replacing, high-quality rice with cheaper rice. This highlights the need for analytical methodologies suitable for its authentication. Given this scenario, the present work aims at testing a rapid and non-destructive approach to detect adulterated rice samples. To fulfill this purpose, 200 rice samples (72 authentic and 128 adulterated samples) were analyzed by near infrared (NIR) spectroscopy coupled, with partial least squares-discriminant analysis (PLS-DA) and soft independent modeling of class analogies (SIMCA). The two approaches provided different results; while PLS-DA analysis was a suitable approach for the purpose of the work, SIMCA was unable to solve the investigated problem. The PLS-DA approach provided satisfactory results in discriminating authentic and adulterated samples (both 5% and 10% counterfeits). Focusing on authentic and 10%-adulterated samples, the accuracy of the approach was even better (with a total classification rate of 82.6% and 82.4%, for authentic and adulterated samples, respectively).

HPLC-DAD fingerprints combined with chemometric techniques for the authentication of plucking seasons of Laoshan green tea

Laoshan green teas plucked in summer and autumn were measured by high performance liquid chromatography-diode array detector (HPLC-DAD). After baseline correction, the fingerprints data were resolved by multivariate curve resolution-alternating least squares (MCR-ALS) and a total of 57 components were acquired. Relative concentrations of these components were afterwards applied to distinguish plucking seasons using principal component analysis (PCA), support vector machines (SVM) and partial least squares-discriminant analysis (PLS-DA). For both SVM and PLS-DA models, the total recognition rates of training set, cross-validation and testing set were 100%, 91.3% and 100%, respectively. Besides, three variable selection methods were employed to determine characteristic components for the authentication of summer and autumn teas. Results showed that PLS-DA model based on three characteristic components selected by VIP possesses identical predictive ability as the original model. This study demonstrated that our proposed strategy is competent for the authentication of plucking seasons of Laoshan green tea.

Exposure to heavy metal contamination and probabilistic health risk assessment using Monte Carlo simulation: a study in the Southeast Iran

The rice contamination to heavy metals and its associated health risks have been less addressed in the southeast of Iran. In the present study, in the mentioned region, we assessed the concentration of nine elements in rice, and the health risk related to the measured elements was determined using the data which were gathered by a questionnaire. For this purpose, 36 samples of the 12 most widely consumed rice brands were collected. Using ICP-MS, the concentrations of Ni, Cr, Hg, Sr, Mn, Fe, Se, Ba, and Zn were measured in the studied samples as 0.599 ± 0.124, 0.483 ± 0.28, 0.0157 ± 0.005, 0.85 ± 1.307, 11.5 ± 1.97, 178.46 ± 67.27, 0.212 ± 0.083, 0.845 ± 0.62, and 8.416 ± 1.611 mg/kg, respectively. We found that, regarding the other studies, the levels of Ni, Cr, Hg, Fe, and Ba were higher. Besides, using 258 distributed questionnaires among citizens, the daily rice consumption was determined to be 295.66 ± 171.005 g/person/ day. Based on this consumption rate and Monto Carlo uncertainty simulation, Fe (0.741 ± 0.54 mg/kg/day) and Se (8.95E-04 ± 6.33E-04 mg/kg/day) showed the highest and lowest daily intake, respectively. Also, using Hazard Quotient (HQ), the non-carcinogenic risk effects of the surveyed elements were estimated. The obtained results of HQ revealed that Fe (2.48) and Mn (1.06) could pose non-carcinogenic health risks to consumers. Moreover, the calculated hazard Index showed that the overall health risk of the surveyed elements is in an unsafe range.

Use of elemental profiles to verify geographical origin and botanical variety of Spanish honeys with a protected denomination of origin

Honey with Protected Denomination of Origin (PDO) could be an attractive target for fraudsters. Elemental profiles by Energy Dispersive-X Ray Fluorescence were processed by multivariate methods to classify 183 PDO honeys produced in three regions of Spain (Liébana, Granada, Tenerife). Additional honey samples (18) produced in a fourth region without PDO (El Bierzo) separated well from the PDO clusters. The manganese content was a discriminant marker of Liébana PDO and El Bierzo, that could also be differentiated from each other. Within each region, distinct clusters revealed differences between dark vs light varieties, multi- vs uni-floral honey and producers of the same PDO. The developed models were validated with 131 samples produced outside the PDO regions and El Bierzo. The proposed classification approach could be implemented as a fast screening tool to support pollen analysis in honey authentication. The reduced number of observations in some light honey models affected their performance.

Comparison of Chemometric Problems in Food Analysis Using Non-Linear Methods

Food analysis is a challenging analytical problem, often addressed using sophisticated laboratory methods that produce large data sets. Linear and non-linear multivariate methods can be used to process these types of datasets and to answer questions such as whether product origin is accurately labeled or whether a product is safe to eat. In this review, we present the application of non-linear methods such as artificial neural networks, support vector machines, self-organizing maps, and multi-layer artificial neural networks in the field of chemometrics related to food analysis. We discuss criteria to determine when non-linear methods are better suited for use instead of traditional methods. The principles of algorithms are described, and examples are presented for solving the problems of exploratory analysis, classification, and prediction.

Long-Term Agricultural Effects on the Authentication Accuracy of Organic, Green, and Conventional Rice Using Isotopic and Elemental Chemometric Analyses

Organically farmed rice is believed to be healthier, safer, and eco-friendlier than its conventionally farmed counterparts and sells for a premium price in global markets. Deliberate mislabeling of organic rice has become a critical consumer concern in China and elsewhere, and there is an increased risk of buying fraudulent organic rice in the market place. In this study, stable isotopic and multielemental analysis combined with chemometrics was used to differentiate organically farmed rice from green and conventional rice in a 4-year experimental field trial from 2014 to 2017. A total of 108 rice samples and their associated soils were collected during the study from three farming (fertilization) systems to investigate whether there are long-term changes in the rice farming classification accuracy from climate effects. Stable carbon and nitrogen isotopic ratios (i.e., δ¹³C and δ¹⁵N) and 27 elemental contents (e.g., Na, K, Ca, Fe, and Zn) of rice and soil samples were determined and then evaluated using statistical analysis [i.e., one-way analysis of variance, multivariable correlation analysis, and modeling of partial least-squares discriminant analysis]. Although δ¹⁵N values can be an effective indicator for organic rice authentication during one crop rotation, both δ¹³C and δ¹⁵N values of rice were easily affected by rice cultivar and interannual soil fertilization and localized agroclimatic variations. These two isotopes were not able to separate organic rice from green and conventional rice accurately. Elemental contents of green and conventional rice (especially K and Ca) were found at higher levels due to the abundant application of synthetic fertilizers (e.g., KNO₃, KH₂PO₄, and CaHPO₄), unlike organically farmed rice, which primarily used animal manure and composts. Partial least-squares discriminant analysis modeling combined isotopic and elemental signatures to correctly differentiate organic rice from green and conventional counterparts, with an accuracy up to 100% over the 4-year study. Therefore, this multi-isotope and -element strategy proposes a more rigorous, alternative tool to combat fraudulent mislabeling of organic rice, increasing the trust of organically labeled rice products and supporting the integrity of the organic sector worldwide.

Air quality warning system based on a localized PM2.5 soft sensor using a novel approach of Bayesian regularized neural network via forward feature selection

It is highly significant to develop efficient soft sensors to estimate the concentration of hazardous pollutants in a region to maintain environmental safety. In this paper, an air quality warning system based on a robust PM_2.5 soft sensor and support vector machine (SVM) classifier is reported. The soft sensor for the estimation of PM_2.5 concentration is proposed using a novel approach of Bayesian regularized neural network (BRNN) via forward feature selection (FFS). Zuoying district of Taiwan is selected as the region of study for implementation of the estimation system because of the high pollution in the region. Descriptive statistics of various pollutants in Zuoying district is computed as part of the study. Moreover, seasonal variation of particulate matter (PM) concentration is analyzed to evaluate the impact of various seasons on the increased levels of PM in the region. To investigate the linear dependence of concentration of different pollutants to the concentration of PM_2.5, Pearson correlation coefficient, Kendall's tau coefficient, and Spearman coefficient are computed. To achieve high performance for the PM_2.5 estimation, selection of appropriate forward features from the input variables is carried out using FFS technique and Bayesian regularization is incorporated to the neural network system to avoid the overfitting problem. The comparative evaluation of performance of BRNN/FFS estimation system with various other methods shows that our proposed estimation system has the lowest mean square error (MSE), root mean square error (RMSE), and mean absolute error (MAE). Moreover, the coefficient of determination (R-squared) is around 0.95 for the proposed estimation method, which denotes a good fit. Evaluation of the SVM classifier showed good performance indicating that the proposed air quality warning system is efficient.

Alternative data mining/machine learning methods for the analytical evaluation of food quality and authenticity - A review

In recent years, the variety and volume of data acquired by modern analytical instruments in order to conduct a better authentication of food has dramatically increased. Several pattern recognition tools have been developed to deal with the large volume and complexity of available trial data. The most widely used methods are principal component analysis (PCA), partial least squares-discriminant analysis (PLS-DA), soft independent modelling by class analogy (SIMCA), k-nearest neighbours (kNN), parallel factor analysis (PARAFAC), and multivariate curve resolution-alternating least squares (MCR-ALS). Nevertheless, there are alternative data treatment methods, such as support vector machine (SVM), classification and regression tree (CART) and random forest (RF), that show a great potential and more advantages compared to conventional ones. In this paper, we explain the background of these methods and review and discuss the reported studies in which these three methods have been applied in the area of food quality and authenticity. In addition, we clarify the technical terminology used in this particular area of research.

Non-essential element concentrations in brown grain rice: Assessment by advanced data mining techniques

The concentrations of 17 non-essential elements (Al, As, Ba, Be, Cd, Ce, Cr, Hg, La, Li, Pb, Sb, Sn, Sr, Th, Ti, and Tl) were determined in brown grain rice samples of two varieties: Fortuna and Largo Fino. The samples were collected from the four main producing regions of Corrientes province (Argentina). Quantitative determinations were performed by inductively coupled plasma mass spectrometry (ICP-MS), using a validated method. The contents of As, Be, Cd, Ce, Cr, Hg, Pb, Sb, Sn, Th, and Tl were very low or not detected in most samples. The non-essential element levels detected were in line with studies conducted in rice from different parts of the world. In order to characterize the influence of geographical origin in the samples, the following classification methods were carried out: linear discriminant analysis (LDA), k-nearest neighbors (k-NN), partial least squares discriminant analysis (PLS-DA), support vector machine (SVM) and random forests (RF). The best performance was obtained by using RF (96%) and SVM (96%). The results reported here showed the variation in the non-essential element profiles in rice grain depending on the geographical origin.

Differentiating Organically Farmed Rice from Conventional and Green Rice Harvested from an Experimental Field Trial Using Stable Isotopes and Multi-Element Chemometrics

Chemometric methods using stable isotopes and elemental fingerprinting were used to characterize organically grown rice from green and conventionally grown rice in experimental field trials in China. Carbon, nitrogen, hydrogen, and oxygen stable isotopes as well as 26 other elements were determined. Organic rice was found to be more depleted in ¹³C than green or conventionally grown rice because of the uptake of enriched ¹³C from carbon dioxide and methane respiring bacteria and more enriched in ¹⁵N because of the volatilization of the nitrogen from the urea and ammonium of the animal manures used to manufacture the organic composts. Chemometrics (principal-component analysis and linear-discriminant analysis) were used to separate the three farming methods and provided a promising scientific tool to authenticate the farming methods of different rice cultivars fertilized with animal manures, green composts, and synthetic fertilizers in China or elsewhere.

Geographic authentication of Asian rice (Oryza sativa L.) using multi-elemental and stable isotopic data combined with multivariate analysis

Rice (Oryza sativa L.) is the world's third largest food crop after wheat and corn. Geographic authentication of rice has recently emerged asan important issue for enhancing human health via food safety and quality assurance. Here, we aimed to discriminate rice of six Asian countries through geographic authentication using combinations of elemental/isotopic composition analysis and chemometric techniques. Principal components analysis could distinguish samples cultivated from most countries, except for those cultivated in the Philippines and Japan. Furthermore, orthogonal projection to latent structure-discriminant analysis provided clear discrimination between rice cultivated in Korea and other countries. The major common variables responsible for differentiation in these models were δ³⁴S, Mn, and Mg. Our findings contribute to understanding the variations of elemental and isotopic compositions in rice depending on geographic origins, and offer valuable insight into the control of fraudulent labeling regarding the geographic origins of rice traded among Asian countries.

Geographical traceability of wild Boletus edulis based on data fusion of FT-MIR and ICP-AES coupled with data mining methods (SVM)

In this work, the data fusion strategy of Fourier transform mid infrared (FT-MIR) spectroscopy and inductively coupled plasma-atomic emission spectrometry (ICP-AES) was used in combination with Support Vector Machine (SVM) to determine the geographic origin of Boletus edulis collected from nine regions of Yunnan Province in China. Firstly, competitive adaptive reweighted sampling (CARS) was used for selecting an optimal combination of key wavenumbers of second derivative FT-MIR spectra, and thirteen elements were sorted with variable importance in projection (VIP) scores. Secondly, thirteen subsets of multi-elements with the best VIP score were generated and each subset was used to fuse with FT-MIR. Finally, the classification models were established by SVM, and the combination of parameter C and γ (gamma) of SVM models was calculated by the approaches of grid search (GS) and genetic algorithm (GA). The results showed that both GS-SVM and GA-SVM models achieved good performances based on the #9 subset and the prediction accuracy in calibration and validation sets of the two models were 81.40% and 90.91%, correspondingly. In conclusion, it indicated that the data fusion strategy of FT-MIR and ICP-AES coupled with the algorithm of SVM can be used as a reliable tool for accurate identification of B. edulis, and it can provide a useful way of thinking for the quality control of edible mushrooms.

Recognition of Mould Colony on Unhulled Paddy Based on Computer Vision using Conventional Machine-learning and Deep Learning Techniques

To investigate the potential of conventional and deep learning techniques to recognize the species and distribution of mould in unhulled paddy, samples were inoculated and cultivated with five species of mould, and sample images were captured. The mould recognition methods were built using support vector machine (SVM), back-propagation neural network (BPNN), convolutional neural network (CNN), and deep belief network (DBN) models. An accuracy rate of 100% was achieved by using the DBN model to identify the mould species in the sample images based on selected colour-histogram parameters, followed by the SVM and BPNN models. A pitch segmentation recognition method combined with different classification models was developed to recognize the mould colony areas in the image. The accuracy rates of the SVM and CNN models for pitch classification were approximately 90% and were higher than those of the BPNN and DBN models. The CNN and DBN models showed quicker calculation speeds for recognizing all of the pitches segmented from a single sample image. Finally, an efficient uniform CNN pitch classification model for all five types of sample images was built. This work compares multiple classification models and provides feasible recognition methods for mouldy unhulled paddy recognition.