András Fekete

Discrimination of mineral waters by electronic tongue, sensory evaluation and chemical analysis.

Mineral, spring and tap water samples of different geographical origins (7 classes) were distinguished by various methods, such as sensory evaluation, electronic tongue measurement, inductively coupled plasma atomic emission spectroscopy and ion chromatography. Samples from the same geographical origin were correctly classified by chemical analysis and electronic tongue (100%), but it was found that only 80% classification rate can be achieved by sensory evaluation. Different water brands (different brand names) from the same geographical origin did not show definite differences, as expected. Forward stepwise algorithm selected three chemical parameters namely, chloride (Cl(-)), sulphate (SO(4)(2-)) and magnesium (Mg) content and two electronic tongue sensor signals (ZZ and HA) to discriminate according to the geographical origins.

Methods for Improving Image Quality and Reducing Data Load of NIR Hyperspectral Images

Near Infrared Hyperspectral Imaging (NIRHSI) is an emerging technology platform that integrates conventional imaging and spectroscopy to attain both spatial and spectral information from an object. Two important problems in NIRHSI are those of data load and unserviceable pixels in the NIR sensor. Hyperspectral imaging experiments generate large amounts of data (typically > 50 MB per image), which tend to overwhelm the memory capacity of conventional computer systems. This inhibits the utilisation of NIRHSI for routine online industrial application. In general, approximately 1% of pixels in NIR detectors are unserviceable or ‘dead’, containing no useful information. While this percentage of pixels is insignificant for single wavelength imaging, the problem is amplified in NIRHSI, where > 100 wavelength images are typically acquired. This paper describes an approach for reducing the data load of hyperspectral experiments by using sample-specific vector-to-scalar operators for real time feature extraction and a systematic procedure for compensating for ‘dead’ pixels in the NIR sensor. The feasibility of this approach was tested for prediction of moisture content in carrot tissue.

Comparison of six multiclass classifiers by the use of different classification performance indicators

Classification problems are very important, and generally, the question is which is the best model. Several classification performance indicators including the classification accuracy value (ACC), Cohens kappa (KAPPA), or the area under the ROC curve (AUC) are used to answer this question. There are non‐parametric comparative methods such as the sum of ranking differences method. The objective of this work was to find the best classification method to classify four soft drink samples and four model samples, which differ from each other only in the sweetener composition. Model samples were used to be basic samples for comparison with the commercial soft drinks. Six different classification methods were compared according to their classification performance. A corrected classification accuracy value (corrected ACC) was developed for the purpose and was introduced. This value takes into account the similarities between the classes. The results showed that the ACC value and the KAPPA values give similar results in our case. The best three models according to the ACC, KAPPA, and AUC were “K‐nearest neighbor,” “random forest,” and “discriminant analysis.” However, the corrected ACC value showed a bit different ranking, and the random forest model was neglected from the good models. The confusion matrices of the models confirmed the ranking according to the corrected ACC value. The results showed that the best classification model was the K‐nearest neighbor for the available samples, and the corrected ACC value is a useful classification performance indicator. Copyright

Sweetener Recognition and Taste Prediction of Coke Drinks by Electronic Tongue

Natural and artificial sweetener monitoring methods are getting more important, since soft drinks with low energy play a considerable role in the market. Our objective is to describe the relevant sensory attributes and to determine the applicability of the electronic tongue to discriminate the coke drink samples with different sweeteners. Furthermore, the aim is to find a relationship between the taste attributes and measurement results received by the electronic tongue. An Alpha astree electronic tongue and a trained sensory panel are used to evaluate coke samples. Panelists found significant differences between the samples in 13 cases from the 18 sensory attributes defined previously by the consensus group. The samples are definitely distinguished by the electronic tongue. The main difference is found according to the sweetener content of the samples. The electronic tongue is able to distinguish samples containing different kinds of artificial and natural sweeteners, as well. The electronic tongue is able to predict, by the partial least squares regression method, the taste attributes of the coke drinks determined by the sensory panel with close correlation and low prediction error.

Mineral Water Taste Attributes Evaluated By Sensory Panel And Electronic Tongue

The objective of the current research was to determine relationships between sensory evaluation and measurement results obtained by electronic tongue for mineral waters. Furthermore, the purpose was to predict the sensory characteristics of the mineral waters measured by the electronic tongue and to determine taste differences that cannot be detected by the sensory evaluation. Two mineral waters were definitely different from the others according to the sensory attributes based on profile analysis. With the electronic tongue measurements the PCA and CDA analysis were found to be able to discriminate mineral waters having chemical composition similar to each other. Very good correlation was found between the sensory attributes and the electronic tongue measurements. However, the results of the measurements performed with the electronic tongue showed a more accurate discrimination of the different mineral waters than the sensory evaluation.

Comparison of Quantitative Determination Techniques with Electronic Tongue Measurements

The objective of the paper reported here was to compare different quantitative techniques for the determination of the adjusted R squares and root mean squared errors of prediction for the different prediction types. The concentration of different solutions was predicted by different quantitative techniques. Linear relationship was found between the logarithm of concentration of different chemical solutions and the sensor response signals. The model of multiple linear regression with the selected sensors gave the highest adjusted R-squared and the lowest root mean squared error of prediction. The root mean squared error of prediction of instant coffee was better than that of mono-sodium glutamate. There was not definite difference between the adjusted R squares for the different prediction types (MLR, PLS and PCR). However, some difference was found between the root mean squared errors of prediction for the different algorithms. Consequently, the electronic tongue was found to be suitable for the analysis of model solutions (e.g. mono-sodium glutamate) and liquid foods (e.g. instant coffee) by the means of the used prediction types and algorithms.

Sensory evaluation and electronic tongue for sensing flavored mineral water taste attributes.

In this article a trained sensory panel evaluated 6 flavored mineral water samples. The samples consisted of 3 different brands, each with 2 flavors (pear-lemon grass and josta berry). The applied sensory method was profile analysis. Our aim was to analyze the sensory profiles and to investigate the similarities between the sensitivity of the trained human panel and an electronic tongue device. Another objective was to demonstrate the possibilities for the prediction of sensory attributes from electronic tongue measurements using a multivariate statistical method (Partial Least Squares regression [PLS]). The results showed that the products manufactured under different brand name but with the same aromas had very similar sensory profiles. The panel performance evaluation showed that it is appropriate (discrimination ability, repeatability, and panel consensus) to compare the panels results with the results of the electronic tongue. The samples can be discriminated by the electronic tongue and an accurate classification model can be built. Principal Component Analysis BiPlot diagrams showed that Brand A and B were similar because the manufacturers use the same aroma brands for their products. It can be concluded that Brand C was quite different compared to the other samples independently of the aroma content. Based on the electronic tongue results good prediction models can be obtained with high correlation coefficient (r(2) > 0.81) and low prediction error (RMSEP < 13.71 on the scale of the sensory evaluation from 0 to 100).

Sensory evaluation and electronic tongue analysis for sweetener recognition in coke drinks

Consumption of beverages with low energy has an increasing role. Furthermore hydrolyzed starch products such as inverted syrup show a wide application in the beverage industry. Therefore the importance of methods which can monitor the usage of natural and artificial sweeteners is increasing. The task was to describe the relevant sensory attributes and to determine the applicability of the electronic tongue to discriminate the coke drink samples with different sweeteners. Furthermore the aim was to find relationship between the taste attributes and measurement results provided by electronic tongue. An Alpha Astree Electronic Tongue and a trained sensory panel were used to evaluate the coke samples. Panelists found significant differences between the samples in 15 cases from the 18 sensory attributes defined previously by the consensus group. Coke drinks containing different kind of sweeteners can be characterized according to these sensory attributes. The samples were definitely distinguished by the electr...

Sensing Basic Tastes by Electronic Tongue Sensors

There is an increasing demand to develop method for simulating the human taste perception by objective instruments1. The task was to develop method for the assessment of definite taste attributes. Therefore, our objective was to develop complete method for sensing different taste attributes. The subject of this work was to test the Specific Sensor Array for taste screening developed by Alpha M.O.S. Different brands of carrot juices were analyzed by an Alpha Astree Electronic Tongue (ET) and a trained sensory panel. The results of the sensory evaluation showed that the different carrot juice samples were significantly different from each other in some taste attributes. The electronic tongue was able to distinguish the tested samples according to the measurement results evaluated by multivariate statistics. Furthermore, the relevant taste attributes of carrot juice samples such as sour taste could be predicted by definite sensors of the electronic tongue. Based on our results we concluded that the selected ...

Taste attributes profiling in carrot juice using an electronic tongue

The taste of five brands of carrot juice was analyzed both by a sensory panel and an electronic tongue. The panelists found significant differences between the carrot juice samples in some appearance and odor attributes and in the relevant taste attributes such as sour taste, sweet taste and taste persistence. Principal component analysis plot calculated from the electronic tongue results showed a clear separation between the sample groups, with a ranking on sourness similar to the one from the panel.