E-nose Data Research Articles

Rapid prediction of the aroma type of plain yogurts via electronic nose combined with machine learning approaches

The unique aroma attributes of yogurts are associated with their quality and consumer acceptance. However, the aroma types of plain yogurts and methods for quickly identifying them have not been systematically established. This study used quantitative descriptive analysis (QDA) to label the aromas of commercial plain yogurts into four primary types, namely fermented, cheesy, milky, and fruity, for the first time. Electronic nose (e-nose) data were rapidly collected from the headspace of plain yogurt samples, and the aroma types classification models were established using linear discriminant analysis (LDA), logistic regression (LR), support vector machine (SVM), and decision trees (DT). Among these models, the SVM-based model displayed the best classification performance, with accuracy and F1 values at 0.933. Furthermore, this study revealed that the LR classification model, based on refined smaller-size e-nose datasets, improved the accuracy and F1 values to 0.867 and 0.872, respectively, compared to the original LR model using the complete dataset, thus demonstrating the applicability of model simplification. This work provides a theoretical basis for rapid plain yogurt aroma type identification by combining e-nose data with machine learning approaches.

A Novel Label Disentangling Subspace Learning Based on Domain Adaptation for Drift E-Nose Data Classification

A Novel Label Disentangling Subspace Learning Based on Domain Adaptation for Drift E-Nose Data Classification

Evaluation of different classification methods using electronic nose data to diagnose sarcoidosis

Electronic nose (eNose) technology is an emerging diagnostic application, using artificial intelligence to classify human breath patterns. These patterns can be used to diagnose medical conditions. Sarcoidosis is an often difficult to diagnose disease, as no standard procedure or conclusive test exists. An accurate diagnostic model based on eNose data could therefore be helpful in clinical decision-making. The aim of this paper is to evaluate the performance of various dimensionality reduction methods and classifiers in order to design an accurate diagnostic model for sarcoidosis. Various methods of dimensionality reduction and multiple hyperparameter optimised classifiers were tested and cross-validated on a dataset of patients with pulmonary sarcoidosis (n = 224) and other interstitial lung disease (n = 317). Best performing methods were selected to create a model to diagnose patients with sarcoidosis. Nested cross-validation was applied to calculate the overall diagnostic performance. A classification model with feature selection and random forest (RF) classifier showed the highest accuracy. The overall diagnostic performance resulted in an accuracy of 87.1% and area-under-the-curve of 91.2%. After comparing different dimensionality reduction methods and classifiers, a highly accurate model to diagnose a patient with sarcoidosis using eNose data was created. The RF classifier and feature selection showed the best performance. The presented systematic approach could also be applied to other eNose datasets to compare methods and select the optimal diagnostic model.

Study on the lipid and flavor compounds of Hibiscus manihot L. oil induced by different pressing method

Hibiscus manihot L. seed oil is an emerging high-value edible vegetable oil that also has a variety of important health functions, and concerns about its pressing method have great significance for commercial and consumer health. This study compared triacylglycerols (TAGs), diacylglycerol (DAG), free fatty acids (FFAs), and volatile flavor compounds for discriminating cold-pressed oil from hot-pressed oil. The profiles of TAGs, DAGs, FFAs, and flavor compounds were established by a lipidomic approach and electronic-nose (E-nose). Our results demonstrated that a slight decrease in TAG content and a significant increase in FFA and DAG contents were observed after roasting. The E-nose results showed that sulfides, terpenes, methane, compounds containing oxygen and nitrogen, alcohol and aromatic compounds greatly influenced the oil flavor. The PCA result of the E-nose data indicated that flavor compounds showed great promise for distinguishing the two samples. The quality indices and the oxidative stability of oil samples were detected to identify their nutritional value and edible quality. This study revealed that roasting greatly influenced the lipid and flavor substances of the Hibiscus manihot L. seed oil. Our findings could provide useful information on the impact of the pressing method for Hibiscus manihot L. seed oil.

Machine olfaction discrimination of Spanish-style green olives inoculated with spoilage mold species

As an alternative to traditional sensory analysis evaluation, machine olfaction represents a valuable tool for detecting incipient defects caused by microorganisms. For this purpose, different strains of spoilage molds were inoculated into Spanish-style table olives and the effect on their sensory quality and volatile profile were analyzed using an electronic nose (E-nose). The main defects obtained for the different inoculated microorganisms associated with abnormal fermentation were mold and humidity. A total of 36 volatile compounds were identified and classified into phenolics, alcohols and carboxylic acids, while minor ones included derivatives of acids and oxygenated compounds. In general, a decrease in the concentrations of acetic acid, phenolic compounds and creosol was observed in olives inoculated with mould strains. However, propanoic acid, 2-methoxy-phenol, 2,4-dimethyl-heptane, among others, increased after mold inoculation. Sensorial and volatile compound analysis showed the table olives inoculated with the strains of Aspergillus flavus A.F.18 and Penicillium expansum P.E.20 to be the most altered. The E-nose data were able to classify the inoculated olives into different categories and distinguish them from the control treatment, regardless the intensity of the defect. The principal component analysis (PCA) of table olives inoculated showed that 70.23% of the total variance of data was explained by PC1 and 23.27% by PC. The sum of the elements of the diagonal of the confusion matrix gives the percentage of success in discrimination of 95.5%. Therefore, these results show the capacity and precision of E-nose to discriminate between the alterations caused by different mold strains.

Determining the shelf life and quality changes of potatoes (Solanum tuberosum) during storage using electronic nose and machine learning.

The activities of alpha-amylase, beta-amylase, sucrose synthase, and invertase enzymes are under the influence of storage conditions and can affect the structure of starch, as well as the sugar content of potatoes, hence altering their quality. Storage in a warehouse is one of the most common and effective methods of storage to maintain the quality of potatoes after their harvest, while preserving their freshness and sweetness. Smart monitoring and evaluation of the quality of potatoes during the storage period could be an effective approach to improve their freshness. This study is aimed at assessing the changes in the potato quality by an electronic nose (e-nose) in terms of the sugar and carbohydrate contents. Three potato cultivars (Agria, Santé, and Sprite) were analyzed and their quality variations were separately assessed. Quality parameters (i.e. sugar and carbohydrate contents) were evaluated in six 15-day periods. The e-nose data were analyzed by means of chemometric methods, including principal component analysis (PCA), linear data analysis (LDA), support vector machine (SVM), and artificial neural network (ANN). Quadratic discriminant analysis (QDA) and multivariate discrimination analysis (MDA) offer the highest accuracy and sensitivity in the classification of data. The accuracy of all methods was higher than 90%. These results could be applied to present a new approach for the assessment of the quality of stored potatoes.

Efficient Monitoring of Microbial Communities and Chemical Characteristics in Incineration Leachate with Electronic Nose and Data Mining Techniques

Incineration leachate is a hazardous liquid waste that requires careful management due to its high levels of organic and inorganic pollutants, and it can have serious environmental and health implications if not properly treated and monitored. This study applied a novel electronic nose to monitor the microbial communities and chemical characteristics of incineration leachate. The e-nose data were aggregated using principal component analysis (PCA) and T-distributed stochastic neighbor embedding (TSNE). Random forest (RF) and gradient-boosted decision tree (GBDT) algorithms were employed to establish relationships between the e-nose signals and the chemical characteristics (such as pH, chemical oxygen demand, and ammonia nitrogen) and microbial communities (including Proteobacteria, Firmicutes, and Bacteroidetes) of the incineration leachate. The PCA-GBDT models performed well in recognizing leachate samples, achieving 100% accuracy for the training set and 98.92% accuracy for the testing data without overfitting. The GBDT models based on the original data performed exceptionally well in predicting changes in chemical parameters, with R2 values exceeding 0.99 for the training set and 0.86 for the testing set. The PCA-GBDT models also demonstrated superior performance in predicting microbial community composition, achieving R2 values above 0.99 and MSE values below 0.0003 for the training set and R2 values exceeding 0.86 and MSE values below 0.015 for the testing set. This research provides an efficient monitoring method for the effective enforcement and implementation of monitoring programs by utilizing e-noses combined with data mining to provide more valuable insights compared with traditional instrumental measurements.

Beef Quality Classification with Reduced E-Nose Data Features According to Beef Cut Types.

Ensuring safe food supplies has recently become a serious problem all over the world. Controlling the quality, spoilage, and standing time for products with a short shelf life is a quite difficult problem. However, electronic noses can make all these controls possible. In this study, which aims to develop a different approach to the solution of this problem, electronic nose data obtained from 12 different beef cuts were classified. In the dataset, there are four classes (1: excellent, 2: good, 3: acceptable, and 4: spoiled) indicating beef quality. The classifications were performed separately for each cut and all cut shapes. The ANOVA method was used to determine the active features in the dataset with data for 12 features. The same classification processes were carried out by using the three active features selected by the ANOVA method. Three different machine learning methods, Artificial Neural Network, K Nearest Neighbor, and Logistic Regression, which are frequently used in the literature, were used in classifications. In the experimental studies, a classification accuracy of 100% was obtained as a result of the classification performed with ANN using the data obtained by combining all the tables in the dataset.

Diagnostic performance of electronic nose technology in chronic lung allograft dysfunction.

There is a need for reliable biomarkers for the diagnosis of chronic lung allograft dysfunction (CLAD). In this light, we investigated the diagnostic value of exhaled breath analysis using an electronic nose (eNose) for CLAD, CLAD phenotype, and CLAD stage in lung transplant recipients (LTR). We performed eNose measurements in LTR with and without CLAD, visiting the outpatient clinic. Through supervised machine learning, the diagnostic value of eNose for CLAD was assessed in a random training and validation set. Next, we investigated the diagnostic value of the eNose measurements combined with known risk factors for CLAD. Model performance was evaluated using ROC-analysis. We included 152 LTR (median age 60 years, 49% females), of whom 38 with CLAD. eNose-based classification of patients with and without CLAD provided an AUC of 0.86 in the training set, and 0.82 in the validation set. After adding established risk factors for CLAD (age, gender, type of transplantation, time after transplantation and prior occurrence of acute cellular rejection) to a model with the eNose data, the discriminative ability of the model improved to an AUC of 0.94 (p=0.02) in the training set and 0.94 (p=0.04) in the validation set. Discrimination between BOS and RAS was good (AUC 0.95). Discriminative ability for other phenotypes (AUCs ranging 0.50-0.92) or CLAD stages (AUC 0.56) was limited. Exhaled breath analysis using eNose is a promising novel biomarker for enabling diagnosis and phenotyping CLAD. eNose technology could be a valuable addition to the diagnostic armamentarium for suspected graft failure in LTR.

Evaluation by a GC Electronic Nose of the Differences in Volatile Profile Induced by Stopping Fermentation with Octanoic and Decanoic Acid to Produce Sweet Wines

Due to their inhibitory effect on the growth and fermentation of yeasts, medium-chain fatty acids can be used for the production of naturally sweet wines. Addition of octanoic acid, decanoic acid or their combinations is able to stop the alcoholic fermentation, reducing at the same time the doses of sulphur dioxide addition needed for the same goal in the classical technologies. Doses in the range of 10–30 mg L−1 of these acids were used, and their effect on the aroma profile of the sweet wines obtained was evaluated by using a chromatographic electronic nose with two columns. Based on the chromatographic peaks, which are considered the sensors of this e-nose, differentiation of the wines treated with octanoic or decanoic acids is easily achieved. The acid doses, the type of acid and also the yeast used for fermentations have all detectable influences on the volatile profiles of the wines. Discriminant factor analysis was applied on the e-nose data to separate the wines obtained with different treatments. Several differences in the content of the volatile compounds were identified and discussed in view of their sensory influences and the impact of treatment and yeast, respectively. Special attention was given to the formation of ethyl octanoate and ethyl decanoate which, at acid additions over 10 mg L−1, are formed in quantities which have a detectable influence on the aromatic profile. Ethyl octanoate and decanoate are produced in direct relation to the dose of the corresponding acids, but the yeast named ST leads to higher amounts of ethyl decanoate while the one named ERSA leads to higher amounts of ethyl octanoate. In accordance with the e-nose results, the aromatic profile obtained by stopping the fermentation with decanoic acid and using the ERSA yeast is more complex, the wines thus produced preserving more of the varietal and fermentation aroma. This research will be continued at an industrial scale.

E-Nose Quality Evaluation of Extra Virgin Olive Oil Stored in Different Containers

The degradation process of virgin olive oil (VOO) is related to storage time and the type of storage container used. The aim of this work is to explore the evolution of the VOO quality stored in different container types over a defined storage period in order to predict the organoleptic characteristics using a non-destructive technique such as the electronic-nose (E-nose). The “Picual” variety VOO was stored in different containers over a period of 21 months and monitored using sensory analysis, volatile compounds, and an E-nose. The panelists showed that oil stored in dark glass bottles and in green polyethylene bottles began to show defects after 12 and 15 weeks, respectively. However, oil stored in tin containers retained its quality throughout the 21 months studied. A total of 31 volatile compounds were identified, and the evolution of the volatile profile in the different containers during the storage period was studied. The E-nose data were able to classify oil quality by container using principal component analysis (PCA). Furthermore, the E-nose data combined with partial least squares (PLS) regression enabled the building of a predictive model to quantify sensory defect values (RCV2 = 0.92; RCV2 = 0.86), evidencing that this technique would be an appropriate screening tool to support a sensory panel.

Combining e-nose and e-tongue for improved recognition of instant starch noodles seasonings.

Seasonings play a key role in determining sensory attributes of instant starch noodles. Controlling and improving the quality of seasoning is becoming important. In this study, five different brands along with fifteen instant starch noodles seasonings (seasoning powder, seasoning mixture sauce and the mixture of powder and sauce) were characterized by electronic nose (e-nose) and electronic tongue (e-tongue). Feature-level fusion for the integration of the signals was introduced to integrate the e-nose and e-tongue signals, aiming at improving the performances of identification and prediction models. Principal component analysis (PCA) explained over 85.00% of the total variance in e-nose data and e-tongue data, discriminated all samples. Multilayer perceptron neural networks analysis (MLPN) modeling demonstrated that the identification rate of the combined data was basically 100%. PCA, cluster analysis (CA), and MLPN proved that the classification results acquired from the combined e-nose and e-tongue data were better than individual e-nose and e-tongue result. This work demonstrated that in combination e-nose and e-tongue provided more comprehensive information about the seasonings compared to each individual e-nose and e-tongue. E-nose and e-tongue technologies hold great potential in the production, quality control, and flavor detection of instant starch noodles seasonings.

FEDA: A Nonlinear Subspace Projection Approach for Electronic Nose Data Classification

The electronic nose (e-nose) is susceptible to sensor drift and instrumental variation, which may result in distribution discrepancy in data collected, hence leading to classification performance degradation. It is necessary to apply domain adaptation to solve this problem. A nonlinear subspace projection approach named feature entropy domain adaption (FEDA) is proposed for domain adaptation for e-nose data classification. One important aspect of FEDA is that adversarial training is introduced to minimize the distribution discrepancy between source and target domains. No projection matrix and parameter fine-tuning are needed anymore in comparison with the popular linear subspace projection approaches. In addition, feature norm and conditional entropy are introduced into adversarial training in FEDA to reduce the decision boundary uncertainty and the overlap between classes, respectively. Experimental results show that the FEDA can deal with the distribution discrepancy of e-nose effectively, and can achieve satisfactory classification accuracy on various datasets. Source code can be found at <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/threedteam/DA_FEDA</uri> .

Analysis of the volatile profiles of kiwifruits experiencing soft rot using E-nose and HS-SPME/GC–MS

Soft rot is a postharvest storage disease occurring in kiwifruit that can cause substantial losses and is difficult to detect. In this study, the internal quality of kiwifruits was measured. In addition, the volatile profiles of kiwifruits were analysed by electronic nose (E-nose) and headspace solid-phase microextraction coupled with gas chromatography–mass spectrometry (HS–SPME/GC–MS). The results showed that the firmness, titratable acidity (TA) and ascorbic acid (AA) decreased as the rot worsened, while malondialdehyde increased. A principal component analysis (PCA) of the E-nose data showed a clear distinction between the intact and diseased fruits. A loading analysis (LOA) of the E-nose data showed that the sensor response appeared to significantly vary between the intact and diseased fruits. Nine differential metabolites for the intact and diseased kiwifruits were acquired by HS–SPME/GC–MS, including isobutyl butyrate, ethyl 2-furoate, ethyl valerianate, propyl butyrate, ethyl hexanoate, 4-terpene alcohols, benzyl alcohol, oxime-methoxy-phenyl and 4-isopropyl toluene. There was a significant correlation (p < 0.05) between the physicochemical indices and volatile profiles obtained by E-nose and HS-SPME/GC–MS. This study proved that the combination of E-nose and GC–MS could potentially be used for detecting kiwifruits experiencing soft rot.

Determination of pitaya quality using portable NIR spectroscopy and innovative low-cost electronic nose

Pitaya (Hylocereus polyrhizus), also known as dragon fruit, is an exotic and highly valued fruit with a high amount of fiber and vitamins, and its quality is often related to attributes such as soluble solids, moisture content, and acidity. Traditional analytical techniques (e.g., gas chromatograph-mass spectrometry – GC–MS) for physicochemical quantification are costly and not environmentally friendly. This work proposes a quick and non-destructive evaluation of pitaya quality using low-cost near-infrared spectroscopy (NIRS) and electronic nose (e-nose) devices. Classification models for either NIR spectra or e-nose data as predictors presented accuracy higher than 90% when classifying samples according to their shelf-life index stage (SLI30, 50, 80, and 100). Total titratable acidity (TA) and pH could be predicted using partial least squares regression (PLSR) and NIR spectra as predictors with coefficients of determination (R2P) of 0.89 and 0.83, respectively, and root means square error (RMSEP) of 0.03 and 0.23, respectively. Similarly, PLSR models for the prediction of TA and pH using e-nose data achieved R2P of 0.85 and 0.86, and RMSEP of 0.04 and 0.22 respectively. RPD and RER values for NIRS show that all predictors can be used to at least distinguish between low and high values. The results demonstrate that inexpensive devices based on NIRS, and a novel low-cost e-nose could be used in combination for the prediction of TSS, pH, TA, moisture, and phenolics, as well as to classify pitaya according to their shelf-life stages.

Qualitative and quantitative assessment of flavor quality of Chinese soybean paste using multiple sensor technologies combined with chemometrics and a data fusion strategy

Multiple sensor technologies including electronic nose (E-nose), electronic tongue (E-tongue), colorimeter and texture analyzer combined with chemometrics and dada fusion strategies were applied to characterize the flavor quality of traditional Chinese fermented soybean paste. Principal components analysis (PCA) was performed to divide the selected soybean pastes into three clusters which was not completely consistent with geographical regions of selected samples. Support vector machine regression (SVR) outperformed partial least squares regression (PLSR) in quantitatively predicting sensory attributes. Additionally, prediction of overall flavor of soybean paste based on data fusion of multiple sensor information, with a correlation coefficient of prediction (Rp) of 0.9636 based on SVR, was better than prediction of E-nose and E-tongue data fusion (Rp = 0.9267). This study suggested multiple sensor technologies coupled with chemometrics can be a promising tool for flavor assessment and characterization of fermented soybean paste or other food matrixes.

Low-Cost Electronic Nose for Wine Variety Identification through Machine Learning Algorithms

The aroma of wine is traditionally analyzed by sensory methods or by using gas chromatography; both analytical methodologies are slow and expensive and do not allow continuous monitoring. For this reason, interest in rapid methods has increased in recent times. Electronic noses (e-noses) stand out for their high sensitivity, speed, low cost, and little or no sample preparation. They present, however, low selectivity, which requires advance analytical methods to distinguish compounds. Here, we present a low-cost e-nose device for the analysis and identification of distinct varieties of wine. Chemical analysis data are compared to e-nose data through a principal component analysis (PCA) and a k-means clustering algorithm to establish relationships between varieties of wines and the e-nose classification capability. The results show that e-nose technology found significant differences between the analyzed samples, and furthermore, classifying the samples in accordance with the chemical analysis classification. The maximal accuracy obtained was 100% using the k-means algorithm for binary classification with N = 21 samples. Thus the potential of e-nose technology was shown in the wine industry for the identification and classification of wine varieties or quality.

Fast and Non-Destructive Profiling of Commercial Coffee Aroma under Three Conditions (Beans, Powder, and Brews) Using GC-IMS.

The flavor of coffee can be affected by the preparation parameters. In this investigation, the flavor profiles of three coffee brands under three conditions (bean, powder, and brew) were analyzed by gas chromatography—ion mobility spectrometry (GC-IMS) and the electronic nose (E-nose). The flavor results were further studied using multiple factor analysis (MFA). A total of 117 peaks were identified in all coffee samples by GC-IMS, and the principal component analysis (PCA) showed these coffee samples could be grouped and separated. A total of 37 volatile organic compounds (VOCs) were selected as biomarkers to distinguish coffee samples, including 5 aldehydes, 10 ketones, 8 alcohols, 2 acids, 4 esters, 5 furans, and 3 other compounds. The comparison between E-nose and GC-IMS data using partial least squares regression (PLSR) and MFA showed GC-IMS could present very close sample spaces. Compared with E-nose, GC–IMS could not only be used to classify coffee samples in a very short time but also provide VOC bio-markers to discriminate coffee samples.

Rapid Assessment of Microbial Quality in Edible Seaweeds Using Sensor Techniques Based on Spectroscopy, Imaging Analysis and Sensors Mimicking Human Senses.

The expansion of the seaweed aquaculture sector along with the rapid deterioration of these products escalates the importance of implementing rapid, real-time techniques for their quality assessment. Seaweed samples originating from Scotland and Ireland were stored under various temperature conditions for specific time intervals. Microbiological analysis was performed throughout storage to assess the total viable counts (TVC), while in parallel FT-IR spectroscopy, multispectral imaging (MSI) and electronic nose (e-nose) analyses were conducted. Machine learning models (partial least square regression (PLS-R)) were developed to assess any correlations between sensor and microbiological data. Microbial counts ranged from 1.8 to 9.5 log CFU/g, while the microbial growth rate was affected by origin, harvest year and storage temperature. The models developed using FT-IR data indicated a good prediction performance on the external test dataset. The model developed by combining data from both origins resulted in satisfactory prediction performance, exhibiting enhanced robustness from being origin unaware towards microbiological population prediction. The results of the model developed with the MSI data indicated a relatively good prediction performance on the external test dataset in spite of the high RMSE values, whereas while using e-nose data from both MI and SAMS, a poor prediction performance of the model was reported.

Nitrogen maceration of wine grape: An alternative and sustainable technique to carbonic maceration

Carbonic maceration (CM) consists in placing intact grape bunches into a sealed tank to have a natural or artificially created carbon dioxide atmosphere. No articles have been published on the comparison between CM and nitrogen maceration (NM). Therefore, the present study aimed at testing the use of alternative maceration technique (NM) in alternative to CM, to create the conditions of anoxia on the Gamay variety. The results showed a higher concentration of polyphenols and anthocyanins in the macerated wines, especially in NM wine. Concerning VOCs (Volatile Organic Compounds), the CM wine showed the highest content in esters while the total alcohol content was slightly higher in the NM wines. The CM wine had the highest content in volatile medium chain fatty acids (MCFA). E-nose measurements revealed a clear separation among the samples, and the E-nose data, in PCA computation, were generally overlapped by the PCA of VOCs analysis. The NM showed a potential as environmentally sustainable technique useful to produce a new style of macerated and aromatic wines.