Discrimination and identification of geographical origin virgin olive oil by an e-nose based on MOS sensors and pattern recognition techniques

Application of Machine Learning Algorithms to Classify Peruvian Pisco Varieties Using an Electronic Nose

Pisco is an alcoholic beverage obtained from grape juice distillation. Considered the flagship drink of Peru, it is produced following strict and specific quality standards. In this work, sensing results for volatile compounds in pisco, obtained with an electronic nose, were analyzed through the application of machine learning algorithms for the differentiation of pisco varieties. This differentiation aids in verifying beverage quality, considering the parameters established in its Designation of Origin". For signal processing, neural networks, multiclass support vector machines and random forest machine learning algorithms were implemented in MATLAB. In addition, data augmentation was performed using a proposed procedure based on interpolation-extrapolation. All algorithms trained with augmented data showed an increase in performance and more reliable predictions compared to those trained with raw data. From the comparison of these results, it was found that the best performance was achieved with neural networks.

Qualitative and Quantitative Detection of Food Adulteration Using a Smart E-Nose

Food adulteration is the most serious problem found in the food industry as it harms people's healths and undermines their beliefs. The present study is focused on designing and developing a smart electronic nose (SE-Nose) for the qualitative and quantitative fast-track detection of food adulteration. The SE-Nose methodology is comprised of a dataset, sample slicing window protocol, normalization, pattern recognition, and output blocks. The dataset pork adulteration in beef is used to validate the SE-Nose methodology. The sample slicing window protocol extracts the early part of the signal. The sample slicing window protocol and pattern recognition models (classification and regression models) together achieved the high-performance and fast-track detection of pork adulteration in beef. With classification models, the qualitative analysis of adulteration is measured, and with regression models, the quantitative analysis of adulteration is measured. An accuracy of 99.996% and an RMSE of 0.02864 were achieved with the SVM classification and regression model. The recognition time in detecting pork adulteration in beef with SVM models is 40 s. With the proposed SE-Nose methodology, the recognition time is reduced by one-third. To validate the classification and regression models, a 10-fold cross-validation method was used.

Volatile Olfactory Profiles of Umbrian Extra Virgin Olive Oils and Their Discrimination through MOX Chemical Sensors

Extra virgin olive oil (EVOO) is the best vegetable oil worldwide but, at the same time, is one of the product victims of fraud in the agri-food sector, and the differences about quality within the extra-virgin olive oil category are often missed. Several scientific techniques were applied in order to guarantee the authenticity and quality of this EVOO. In the present study, the volatile compounds (VOCs) by gas chromatography-mass spectrometry with solid-phase micro-extraction detection (GC-MS SPME), organoleptic analysis by the official Slow Food panel and the detection by a Small Sensor System (S3) were applied. Ten EVOOs from Umbria, a central Italian region, were selected from the 2021 Slow Food Italian extra virgin olive oil official guide, which includes hundreds of high-quality olive oils. The results demonstrated the possibility to discriminate the ten EVOOs, even if they belong to the same Italian region, by all three techniques. The result of GC-MS SPME detection was comparable at the discrimination level to the organoleptic test with few exceptions, while the S3 was able to better separate some EVOOs, which were not discriminated perfectly by the other two methods. The correlation analysis performed among and between the three methodologies allowed us to identify 388 strong associations with a p value less than 0.05. This study has highlighted how much the mix of VOCs was different even among few and localized EVOOs. The correlation with the sensor detection, which is faster and chipper compared to the other two techniques, elucidated the similarities and discrepancies between the applied methods.

Advanced techniques in edible oil authentication: A systematic review and critical analysis

Adulteration of edible substances is a potent contemporary food safety issue. Perhaps the overt concern derives from the fact that adulterants pose serious ill effects on human health. Edible oils are one of the most adulterated food products. Perpetrators are adopting ways and means that effectively masks the presence of the adulterants from human organoleptic limits and traditional oil adulteration detection techniques. This review embodies a detailed account of chemical, biosensors, chromatography, spectroscopy, differential scanning calorimetry, non-thermal plasma, dielectric spectroscopy research carried out in the area of falsification assessment of edible oils for the past three decades and a collection of patented oil adulteration detection techniques. The detection techniques reviewed have some advantages and certain limitations, chemical tests are simple; biosensors and nuclear magnetic resonance are rapid but have a low sensitivity; chromatography and spectroscopy are highly accurate with a deterring price tag; dielectric spectroscopy is rapid can be portable and has on-line compatibility; however, the results are susceptible to variation of electric current frequency and intrinsic factors (moisture, temperature, structural composition). This review paper can be useful for scientists or for knowledge seekers eager to be abreast with edible oil adulteration detection techniques.

An Outlook of Recent Advances in Chemiresistive Sensor-Based Electronic Nose Systems for Food Quality and Environmental Monitoring

An electronic nose (Enose) relies on the use of an array of partially selective chemical gas sensors for identification of various chemical compounds, including volatile organic compounds in gas mixtures. They have been proposed as a portable low-cost technology to analyse complex odours in the food industry and for environmental monitoring. Recent advances in nanofabrication, sensor and microcircuitry design, neural networks, and system integration have considerably improved the efficacy of Enose devices. Here, we highlight different types of semiconducting metal oxides as well as their sensing mechanism and integration into Enose systems, including different pattern recognition techniques employed for data analysis. We offer a critical perspective of state-of-the-art commercial and custom-made Enoses, identifying current challenges for the broader uptake and use of Enose systems in a variety of applications.

Application of a lab-made electronic nose for extra virgin olive oils commercial classification according to the perceived fruitiness intensity

An electronic nose, comprising nine metal oxide sensors, has been built aiming to classify olive oils according to the fruity intensity commercial grade (ripely fruity or light, medium and intense greenly fruity), following the European regulated complementary terminology. The lab-made sensor device was capable to differentiate standard aqueous solutions (acetic acid, cis-3-hexenyl, cis-3-hexen-1-ol, hexanal, 1-hexenol and nonanal) that mimicked positive sensations (e.g., fatty, floral, fruit, grass, green and green leaves attributes) and negative attributes (e.g., sour and vinegary defects), as well as to semi-quantitatively classify them according to the concentration ranges (0.05-2.25 mg/kg). For that, unsupervised (principal component analysis) and supervised (linear discriminant analysis: sensitivity of 92% for leave-one-out cross validation) classification multivariate models were established based on nine or six gas sensors, respectively. It was also showed that the built E-nose allowed differentiating/discriminating (sensitivity of 81% for leave-one-out cross validation) extra virgin olive oils according to the perceived intensity of fruitiness as ripely fruity, light, medium or intense greenly fruity. In conclusion, the gas sensor device could be used as a practical preliminary non-destructive tool for guaranteeing the correctness of olive oil fruitiness intensity labelling.

Biodegradable Food Packaging Materials and Prospects of the Fourth Industrial Revolution for Tomato Fruit and Product Handling

The environment and food safety are major areas of concern influencing the development of biodegradable packaging for partial replacement of petrochemical-based polymers. This review is aimed at updating the recent advances in biodegradable packaging material and the role of virtual technology and nanotechnology in the tomato supply chain. Some of the common biodegradable materials are gelatin, starch, chitosan, cellulose, and polylactic acid. The tensile strength, tear resistance, permeability, degradability, and solubility are some of the properties defining the selection and utilization of food packaging materials. Biodegradable films can be degraded in soil by microbial enzymatic actions and bioassimilation. Nanoparticles are incorporated into blended films to improve the performance of packaging materials. The prospects of the fourth industrial revolution can be realized with the use of virtual platforms such as sensor systems in authentification and traceability of food and packaging products. There is a research gap on the development of a hybrid sensor system unit that can integrate sampling headspace (SHS), detection unit, and data processing of big data for heterogeneous tomato-derived volatiles. Principal component analysis (PCA), linear discriminant analysis (LDA), and artificial neutral network (ANN) are some of the common mathematical models for data interpretation of sensor systems.

Multidisciplinary Approach to Characterizing the Fingerprint of Italian EVOO

Extra virgin olive oil (EVOO) is characterized by its aroma and other sensory attributes. These are determined by the geographical origin of the oil, extraction process, place of cultivation, soil, tree varieties, and storage conditions. In the present work, an array of metal oxide gas sensors (called S3), in combination with the SPME-GC-MS technique, was applied to the discrimination of different types of olive oil (phase 1) and to the identification of four varieties of Garda PDO extra virgin olive oils coming from west and east shores of Lake Garda (phase 2). The chemical analysis method involving SPME-GC-MS provided a complete volatile component of the extra virgin olive oils that was used to relate to the S3 multisensory responses. Furthermore, principal component analysis (PCA) and k-Nearest Neighbors (k-NN) analysis were carried out on the set of data acquired from the sensor array to determine the best sensors for these tasks and to assess the capability of the system to identify various olive oil samples. k-NN classification rates were found to be 94.3% and 94.7% in the two phases, respectively. These first results are encouraging and show a good capability of the S3 instrument to distinguish different oil samples.

Electrochemical Sensor-Based Devices for Assessing Bioactive Compounds in Olive Oils: A Brief Review

Electrochemical bioinspired sensor devices combined with chemometric tools have experienced great advances in the last years, being extensively used for food qualitative and quantitative evaluation, namely for olive oil analysis. Olive oil plays a key role in the Mediterranean diet, possessing unique and recognized nutritional and health properties as well as highly appreciated organoleptic characteristics. These positive attributes are mainly due to olive oil richness in bioactive compounds such as phenolic compounds. In addition, these compounds enhance their overall sensory quality, being mainly responsible for the usual olive oil pungency and bitterness. This review aims to compile and discuss the main research advances reported in the literature regarding the use of electrochemical sensor based-devices for assessing bioactive compounds in olive oil. The main advantages and limitations of these fast, accurate, bioinspired voltammetric, potentiometric and/or amperometric sensor green-approaches will be addressed, aiming to establish the future challenges for becoming a practical quality analytical tool for industrial and commercial applications.

An overview of an artificial nose system

The present review describes recent advances in the development of an artificial nose system based on olfactory receptors and various sensing platforms. The kind of artificial nose, the production of olfactory receptors, the sensor platform for signal conversion and the application of the artificial nose system based on olfactory receptors and various sensing platforms are presented. The associated transduction modes are also discussed. The paper presents a review of the latest achievements and a critical evaluation of the state of the art in the field of artificial nose systems.

Artificial nose, NIR and UV-visible spectroscopy for the characterisation of the PDO Chianti Classico olive oil

An authentication study of the Italian PDO (Protected Designation of Origin) olive oil Chianti Classico, based on artificial nose, near-infrared and UV-visible spectroscopy, with a set of samples representative of the whole Chianti Classico production area and a considerable number of samples from other Italian PDO regions was performed. The signals provided by the three analytical techniques were used both individually and jointly, after fusion of the respective variables, in order to build a model for the Chianti Classico PDO olive oil. Different signal pre-treatments were performed in order to investigate their importance and their effects in enhancing and extracting information from experimental data, correcting backgrounds or removing baseline variations. Stepwise-Linear Discriminant Analysis (STEP-LDA) was used as a feature selection technique and, afterward, Linear Discriminant Analysis (LDA) and the class-modelling technique Quadratic Discriminant Analysis-UNEQual dispersed classes (QDA-UNEQ) were applied to sub-sets of selected variables, in order to obtain efficient models capable of characterising the extra virgin olive oils produced in the Chianti Classico PDO area.