A critical review on the applications of artificial neural networks in winemaking technology

Rapid classification of Camellia seed varieties and non-destructive high-throughput quantitative analysis of fatty acids based on non-targeted fingerprint spectroscopy combined with chemometrics

Camellia oil is a high-quality vegetable oil rich in unsaturated fatty acids (FAs), with quality standardization challenged by the diversity of Camellia seed varieties. This study compared spectroscopy techniques (Near-Infrared [NIR] vs Mid-Infrared [MIR] spectroscopy) and analytical models (Discriminant Analysis [DA], Partial Least Squares [PLS], and Artificial Neural Networks [ANN]), seeking to classify Camellia seed varieties and estimate oil and principal FAs composition. The PCA analysis effectively discriminated among various Camellia seed varieties, likely due to variations in their oil and principal FAs compositions. Significantly, the NIR-based DA model significantly outperformed MIR, achieving 100 % accuracy in distinguishing Camellia seed varieties. In terms of predicting the oil and principal FAs compositions in Camellia seeds, NIR-based predictions models outperformed those derived from MIR, with PLS models surpassing ANN models. This study validated the potential of NIR technology combined with chemometrics for rapid, high-throughput, non-destructive identification of Camellia seeds.

Optimizing Recovery of High-Added-Value Compounds from Complex Food Matrices Using Multivariate Methods

In today's food industry, optimizing the recovery of high-value compounds is crucial for enhancing quality and yield. Multivariate methods like Response Surface Methodology (RSM) and Artificial Neural Networks (ANNs) play key roles in achieving this. This review compares their technical strengths and examines their sustainability impacts, highlighting how these methods support greener food processing by optimizing resources and reducing waste. RSM is valued for its structured approach to modeling complex processes, while ANNs excel in handling nonlinear relationships and large datasets. Combining RSM and ANNs offers a powerful, synergistic approach to improving predictive models, helping to preserve nutrients and extend shelf life. The review emphasizes the potential of RSM and ANNs to drive innovation and sustainability in the food industry, with further exploration needed for scalability and integration with emerging technologies.

Machine Learning Models to Classify Shiitake Mushrooms (Lentinula edodes) According to Their Geographical Origin Labeling

The shiitake mushroom has gained popularity in the last decade, ranking second in the world for mushrooms consumed, providing consumers with a wide variety of nutritional and healthy benefits. It is often not clear the origin of these mushrooms, so it becomes of great importance to the consumers. In this research, different machine learning algorithms were developed to determine the geographical origin of shiitake mushrooms (Lentinula edodes) consumed in Korea, based on experimental data reported in the literature (δ13C, δ15N, δ18O, δ34S, and origin). Regarding the origin of shiitake in three categories (Korean, Chinese, and mushrooms from Chinese inoculated sawdust blocks), the random forest model presents the highest accuracy value (0.940) and the highest kappa value (0.908) for the validation phase. To determine the origin of shiitake mushrooms in two categories (Korean and Chinese, including mushrooms from Chinese inoculated sawdust blocks in the latter ones), the support vector machine model is chosen as the best model due to the high accuracy (0.988) and kappa (0.975) values for the validation phase. Finally, to determine the origin in two categories (Korean and Chinese, but this time including the mushrooms from Chinese inoculated sawdust blocks in the Korean ones), the best model is the random forest due to its higher accuracy value (0.952) in the validation phase (kappa value of 0.869). The accuracy values in the testing phase for the best selected models are acceptable (between 0.839 and 0.964); therefore, the predictive capacity of the models could be acceptable for their use in real applications. This allows us to affirm that machine learning algorithms would be suitable modeling instruments to determine the geographical origin of shiitake.

Modernization of digital food safety control

Foodborne illness remains a pressing global issue due to the complexities of modern food supply chains and the vast array of potential contaminants that can arise at every stage of food processing from farm to fork. Traditional food safety control systems are increasingly challenged to identify these intricate hazards. The U.S. Food and Drug Administration's (FDA) New Era of Smarter Food Safety represents a revolutionary shift in food safety methodology by leveraging cutting-edge digital technologies. Digital food safety control systems employ modern solutions to monitor food quality by efficiently detecting in real time a wide range of contaminants across diverse food matrices within a short timeframe. These systems also utilize digital tools for data analysis, providing highly predictive assessments of food safety risks. In addition, digital food safety systems can deliver a secure and reliable food supply chain with comprehensive traceability, safeguarding public health through innovative technological approaches. By utilizing new digital food safety methods, food safety authorities and businesses can establish an efficient regulatory framework that genuinely ensures food safety. These cutting-edge approaches, when applied throughout the food chain, enable the delivery of safe, contaminant-free food products to consumers.

An efficient methodology for modeling to predict wine aroma expression based on quantitative data of volatile compounds: A case study of oak barrel-aged red wines

Correlating aroma expression with volatile compounds has long been an ambition in researches of flavor chemistry. To propose a reliable methodology to depict wine aroma, 76 oak barrel-aged dry red wines were investigated through the combination of machine learning algorithm and multivariate analysis. Aromatic characteristic was evaluated by quantitative descriptive analysis (QDA), while non- or oak derived volatiles were detected by HS-SPME-GC-MS and targeted SPE-GC-QqQ-MS/MS, respectively. Results showed that variable importance for projection values (VIPs) from partial least-squares regression (PLSR) and mean decrease accuracy (MDA) from random forest were efficient parameters for feature selection. The correlating accuracy of the optimal PLSR model to predict intensities of different aroma characteristics through selected volatile compounds could achieve 0.754 to 0.943, representing potential application to manage wine aroma by chemical assay in winemaking. From the perspective of mathematical modeling in the real wine matrix, the network analysis between aroma characteristics and key volatile compounds indicated that the expression of oak aroma was not only directly contributed by volatiles derived from oak wood, but also influenced by ethyl esters, including ethyl acetate, ethyl butanoate, ethyl hexanoate, ethyl decanoate, and ethyl nonanoate.

Co-winemaking with Vitis amurensis Rupr. "Beibinghong" enhances the quality of Vitis vinifera L. cv. Cabernet Gernischt wine

In some wine regions of China, Cabernet Gernischt (CG; Vitis vinifera L.) grape berries usually exhibit low pigment content and titratable acidity, and low sensory quality of the resulting wine. The aim of this study was to evaluate co-winemaking of CG wines using the red grape cultivar Beibinghong (BBH; Vitis amurensis Rupr.) at different proportions in terms of alcohols, phenolic compounds, and sensory properties of the wines. The results showed that the co-winemaking wines contained a similar content of higher alcohols, whereas the methanol content increased with an increase in BBH proportion, although this was still corresponded with the national standard. Significantly higher levels of titratable acidity were observed in co-winemaking wines at the ratio of 6:4 and 5:5, compared with monocultivar CG wines. All co-winemaking wines, except CG:BBH (9:1) wine, showed significantly higher levels of total anthocyanins, total phenolics, total tannins, and total flavan-3-ols. Further, individual phenolics, primarily diglucoside anthocyanins and non-anthocyanins (trans-ferulic acid, myricetin, viniferin, trans-caffeic acid, 3,4-dihydroxybenzoic acid), as important contributors to wine color intensity, permitted the differentiation of the wines via principal component analysis. In most cases, co-winemaking wines exhibited higher scores of the 10 sensory attributes on color, aroma, mouthfeel, and overall quality compared with monocultivar wines. Co-winemaking CG wines with BBH at 7:3 ratio demonstrated the highest scores of color intensity, aroma intensity, aroma quality, and overall quality. The results indicate that co-winemaking with V. amurensis grape variety may be useful to enhance V. vinifera wine quality by modifying wine composition. PRACTICAL APPLICATION: Cabernet Gernischt (CG) is the predominant grape cultivar used to prepare premium-quality wine in China; however, in some wine regions, CG wines have low levels of pigment and titratable acidity, and low sensory quality. Co-winemaking with another native grape cultivar, Beibinghong (BBH), which is characterized by a higher content of anthocyanins and acidity, provided sufficient experimental evidence of adjustments in the Vitis vinifera wine composition leading to improved wine sensory quality.

Fourier transform infrared spectroscopy in monitoring the wine production

The complexity of the wine matrix makes monitoring of the winemaking process from the grapes to the final product crucial for the wine industry. In this context, analytical methodologies that can combine good accuracy, robustness, high sample throughput, “green character”, and by preference real-time analysis, are on-demand to create high-quality vitivinicultural products. In the last years, Fourier-transform Infrared Spectroscopy (FTIR) combined with chemometric analysis has been evaluated in several studies as an effective analytical tool for the wine sector. Some applications of FTIR spectroscopy have been already accepted by the wine industry, mainly for the prediction of basic oenological parameters, using portable and non-portable instruments, but still many others are waiting to be thoroughly developed. This literature review aims to provide a critical synopsis of the most important studies assessing grape and wine quality and authenticity, and to identify possible gaps for further research, meeting the needs of the modern wine industry and the expectations of most demanding consumers. The FTIR studies were grouped according to the main sampling material used - 1) leaves, stems, and berries; 2) grape must and wine applications - along with a summary of the basic limitations and future perspectives of this analytical technique.

Use of Multivariate Statistics in the Processing of Data on Wine Volatile Compounds Obtained by HS-SPME-GC-MS

This review takes a snapshot of the main multivariate statistical techniques and methods used to process data on the concentrations of wine volatile molecules extracted by means of solid phase micro-extraction and analyzed using GC-MS. Hypothesis test, exploratory analysis, regression models, and unsupervised and supervised pattern recognition methods are illustrated and discussed. Several applications in the wine volatolomic sector are described to highlight different interactions among the various matrix components and volatiles. In addition, the use of Artificial Intelligence-based methods is discussed as an innovative class of methods for validating wine varietal authenticity and geographical traceability.

Physicochemical parameters combined flash GC e-nose and artificial neural network for quality and volatile characterization of vinegar with different brewing techniques

Vinegar is a kind of traditional fermented food, there are significant variances in quality and flavor due to differences in raw ingredients and processes. The quality assessment and flavor characteristics of 69 vinegar samples with 5 brewing processes were analyzed by physicochemical parameters combined with flash gas chromatography (GC) e-nose. The evaluation system of quality and the detection method of flavor profile were established. 17 volatile flavor compounds and potential flavor differential compounds of each brewing process were identified. The artificial neural network (ANN) analysis model was established based on the physicochemical parameters and the analysis of flash GC e-nose. Although the physicochemical parameters were more intuitive in quality evaluating, the flash GC e-nose could better reflect the flavor characteristics of vinegar samples and had better fitting, prediction and discrimination ability, the correct rates of training and prediction of flash GC e-nose trained ANN model were 98.6% and 96.7%, respectively.

Detection of fraud in lime juice using pattern recognition techniques and FT-IR spectroscopy

The lime juice is one of the products that has always fallen victim to fraud by manufacturers for reducing the cost of products. The aim of this research was to determine fraud in distributed lime juice products from different factories in Iran. In this study, 101 samples were collected from markets and also prepared manually and finally derived into 5 classes as follows: two natural classes (Citrus limetta, Citrus aurantifolia), including 17 samples, and three reconstructed classes, including 84 samples (made from Spanish concentrate, Chinese concentrate, and concentrate containing adulteration compounds). The lime juice samples were freeze-dried and analyzed using FT-IR spectroscopy. At first, principal component analysis (PCA) was applied for clustering, but the samples were not thoroughly clustered with respect to their original groups in score plots. To enhance the classification rates, different chemometric algorithms including variable importance in projection (VIP), partial least square-discriminant analysis (PLS-DA), and counter propagation artificial neural networks (CPANN) were used. The best discriminatory wavenumbers related to each class were selected using the VIP-PLS-DA algorithm. Then, the CPANN algorithm was used as a nonlinear mapping tool for classification of the samples based on their original groups. The lime juice samples were correctly designated to their original groups in CPANN maps and the overall accuracy of the model reached up to 0.96 and 0.87 for the training and validation procedures. This level of accuracy indicated the FT-IR spectroscopy coupled with VIP-PLS-DA and CPANN methods can be used successfully for detection of authenticity of lime juice samples.

Modification of Sensory Expression of 3-Isobutyl-2-methoxypyrazine in Wines through Blending Technique

Sensory interactions exist between 3-alkyl-2-methoxypyrazines and various volatiles in wines. In this study, the binary blending of Cabernet Franc wines containing high levels of MPs and three monovarietal red wines with two proportions was conducted after fermentation. Volatiles were detected by gas chromatography-mass spectrometry (GC-MS), and wines were evaluated by quantitative descriptive analysis at three-month intervals during six-month bottle aging. Results showed blending wines exhibited lower intensity of ‘green pepper’, especially CFC samples blended by Cabernet Sauvignon wines with an even higher concentration of 3-isobutyl-2-methoxypyrazine (IBMP). Based on Pearson correlation analysis, acetates could promote the expression of ‘tropical fruity’ and suppress ‘green pepper’ caused by IBMP. Positive correlation was observed among ‘green pepper’, ‘herbaceous’, and ‘berry’. The concentration balance between IBMP and other volatiles associated with ‘green pepper’ and fruity notes was further investigated through sensory experiments in aroma reconstitution. Higher pleasant fruity perception was obtained with the concentration proportion of 1-hexanol (1000 μg/L), isoamyl acetate (550 μg/L), ethyl hexanoate (400 μg/L), and ethyl octanoate (900 μg/L) as in CFC samples. Blending wines with proper concentration of those volatiles would be efficient to weaken ‘green pepper’ and highlight fruity notes, which provided scientific theory on sensory modification of IBMP through blending technique.

Quality control method of sterols in fermented Cordyceps sinensis based on combined fingerprint and quantitative analysis of multicomponents by single marker

In this study, we established a new pattern for differentiating and comprehensively evaluating the quality of fermented Cordyceps sinensis based on high-performance liquid chromatography (HPLC) fingerprint analysis combined with similar analysis (SA), principal component analysis (PCA), hierarchical cluster analysis (HCA), and the quantitative analysis of multicomponents by single marker (QAMS). These methods indicated that fermented Cordyceps sinensis samples could be categorized into one class by PCA and HCA. The fingerprints of fermented Cordyceps sinensis were established, and four HPLC peaks were identified as ergosterol, daucosterol, stigmasterol, and β-sitosterol in Jinshuibao capsules and tablets (two products of fermented Cordyceps sinensis). Ergosterol was chosen as the internal reference substance, and the relative correction factors (RCFs) between ergosterol and the other three sterols were calculated using the QAMS method. Moreover, the accuracy of the QAMS method was confirmed by comparing the relative error between the results of the method used with those of an external standard method (ESM). No significant difference between the two methods was observed. The total sterols content in Jinshuibao products were calculated by the QAMS method, and the total sterols content of the two products were similar. This study showed that the method established herein was efficient and successful in the identification fermented Cordyceps sinensis and may further act to facilitate systematic quality control of fermented Cordyceps sinensis products.

Electrochemical Sensors Coupled with Multivariate Statistical Analysis as Screening Tools for Wine Authentication Issues: A Review

Consumers are increasingly interested in the characteristics of the products they consume, including aroma, taste, and appearance, and hence, scientific research was conducted in order to develop electronic senses devices that mimic the human senses. Thanks to the utilization of electroanalytical techniques that used various sensors modified with different electroactive materials coupled with pattern recognition methods, artificial senses such as electronic tongues (ETs) are widely applied in food analysis for quality and authenticity approaches. This paper summarizes the applications of electrochemical sensors (voltammetric, amperometric, and potentiometric) coupled with unsupervised and supervised pattern recognition methods (principal components analysis (PCA), linear discriminant analysis (LDA), partial least square (PLS) regression, artificial neural network (ANN)) for wine authenticity assessments including the discrimination of varietal and geographical origins, monitoring the ageing processes, vintage year discrimination, and detection of frauds and adulterations. Different wine electrochemical authentication methodologies covering the electrochemical techniques, electrodes types, functionalization sensitive materials and multivariate statistical analysis are emphasized and the main advantages and disadvantages of using the proposed methodologies for real applications were concluded.

Variation in secondary metabolites in a unique set of tomato accessions collected in Turkey

In this study, 50 tomato landraces grown in Turkey were investigated in terms of their secondary metabolite profiles. Each accession was planted in 2016 and 2017 in 3 replicates in an open field. In this study, color, pH and brix of the fruit samples were measured and an unbiased LCMS-based metabolomics approach was applied. Based on Principal Components Analysis (PCA) and Hierarchical Cluster Analysis (HCA) of the relative abundance levels of >250 metabolites, it could be concluded that fruit size was the most influential to the biochemical composition, rather than the geographical origin of accessions. Results indicated substantial biodiversity in various metabolites generally regarded as key to fruit quality aspects, including sugars; phenolic compounds like phenylpropanoids and flavonoids; alkaloids and glycosides of flavour-related volatile compounds. The phytochemical data provides insight into which Turkish accessions might be most promising as starting materials for the tomato processing and breeding industries.

Prediction Models to Control Aging Time in Red Wine

A combination of physical-chemical analysis has been used to monitor the aging of red wines from D.O. Toro (Spain). The changes in the chemical composition of wines that occur over the aging time can be used to distinguish between wine samples collected after one, four, seven and ten months of aging. Different computational models were used to develop a good authenticity tool to certify wines. In this research, different models have been developed: Artificial Neural Network models (ANNs), Support Vector Machine (SVM) and Random Forest (RF) models. The results obtained for the ANN model developed with sigmoidal function in the output neuron and the RF model permit us to determine the aging time, with an average absolute percentage deviation below 1%, so it can be concluded that these two models have demonstrated their capacity to predict the age of wine.

Fatty Acids-Based Quality Index to Differentiate Worldwide Commercial Pistachio Cultivars

The fatty acid profiles of five main commercial pistachio cultivars, including Ahmad-Aghaei, Akbari, Chrok, Kalle-Ghouchi, and Ohadi, were determined by gas chromatography: palmitic (C16:0), palmitoleic (C16:1), stearic (C18:0), oleic (C18:1), linoleic (C18:2), linolenic (C18:3), arachidic (C20:0), and gondoic (C20:1) acid. Based on the oleic to linoleic acid (O/L) ratio, a quality index was determined for these five cultivars: Ohadi (2.40) < Ahmad-Aghaei (2.60) < Kale-Ghouchi (2.94) < Chrok (3.05) < Akbari (3.66). Principal component analysis (PCA) of the fatty acid data yielded three significant PCs, which together account for 80.0% of the total variance in the dataset. A linear discriminant analysis (LDA) model that was evaluated with cross-validation correctly classified almost all of the samples: the average percent accuracy for the prediction set was 98.0%. The high predictive power for the prediction set shows the ability to indicate the cultivar of an unknown sample based on its fatty acid chromatographic fingerprint.