Neural networks applied to discriminate botanical origin of honeys

Anjos, Ofélia; Iglesias, Carla; Péres, Fernando Curi; Martínez, Javier Martínez; García, Ángela; Taboada, J.

doi:10.1016/j.foodchem.2014.11.121

Cited by 80 publications

(63 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Honey carbohydrates are composed up of about 70% monosaccharides (mainly glucose and fructose), 10-15% disaccharides and a minor concentration of trisaccharides [1]. The chemical composition of this beehive product depends on the botanical and geographical origin [3], which may be evaluated through several methodologies.…”

Section: Introductionmentioning

confidence: 99%

Physicochemical characterization of Lavandula spp. honey with FT-Raman spectroscopy

Anjos

Santos

Paixão³

et al. 2018

Talanta

Self Cite

View full text Add to dashboard Cite

A B S T R A C TThis study aimed to evaluate the potential of FT-Raman spectroscopy in the prediction of the chemical composition of Lavandula spp. monofloral honey. Partial Least Squares (PLS) regression models were performed for the quantitative estimation and the results were correlated with those obtained using reference methods.Good calibration models were obtained for electrical conductivity, ash, total acidity, pH, reducing sugars, hydroxymethylfurfural (HMF), proline, diastase index, apparent sucrose, total flavonoids content and total phenol content. On the other hand, the model was less accurate for pH determination. The calibration models had high r 2 (ranging between 92.8% and 99.9%), high residual prediction deviation -RPD (ranging between 4.2 and 26.8) and low root mean square errors. These results confirm the hypothesis that FT-Raman is a useful technique for the quality control and chemical properties' evaluation of Lavandula spp honey. Its application may allow improving the efficiency, speed and cost of the current laboratory analysis.

show abstract

Section: Introductionmentioning

confidence: 99%

Physicochemical characterization of Lavandula spp. honey with FT-Raman spectroscopy

Anjos

Santos

Paixão³

et al. 2018

Talanta

Self Cite

View full text Add to dashboard Cite

show abstract

“…Computational intelligence models based on NN, FM and SVM focuses on the biological learning process and try to emulate it through algorithms able to learn from given data and provide new results (Anjos et al, 2015). These techniques have been claimed to be an effective tool for modelling and predicting different parameters among which the rheological behavior of honey (Ramzi, Kashaninejad, Salehi, Sadeghi Mahoonak, & Ali Razavi, 2015).…”

Section: Introductionmentioning

confidence: 99%

“…Also, these tools were also applied in the identification of the relationship between honeys' chemical and electrical parameters (Pentoś, Łuczycka, & Wróbel, 2014). Anjos et al (2015) using neural network, conclude that the botanical origin of honey can be reliably and quickly known from the colorimetric information and the electrical conductivity of honey. Shafiee, Minaei, Moghaddam-Charkari, GhasemiVarnamkhasti, and Barzegar (2013) antioxidant activity and total phenolic content from honey, using neural networks and color information.…”

Section: Introductionmentioning

confidence: 99%

Computational intelligence applied to discriminate bee pollen quality and botanical origin

et al. 2018

Self Cite

View full text Add to dashboard Cite

a b s t r a c tThe aim of this work was to develop computational intelligence models based on neural networks (NN), fuzzy models (FM), and support vector machines (SVM) to predict physicochemical composition of bee pollen mixture given their botanical origin. To obtain the predominant plant genus of pollen (was the output variable), based on physicochemical composition (were the input variables of the predictive model), prediction models were learned from data. For the inverse case study, input/output variables were swapped.The probabilistic NN prediction model obtained 98.4% of correct classification of the predominant plant genus of pollen. To obtain the secondary and tertiary plant genus of pollen, the results present a lower accuracy.To predict the physicochemical characteristic of a mixture of bee pollen, given their botanical origin, fuzzy models proven the best results with small prediction errors, and variability lower than 10%.

show abstract

“…An excellent separation (100%) among honey samples according to their botanical origin in MLP model with amino acids content as input parameters was presented by Chen et al [20]. About 95% accuracy was achieved by Anjos et al [21] when MLP classified honey samples according to botanical origin on the basis of the colorimetric information and the electrical conductivity. MLP was adopted by Zhu et al [22] to correctly (90.2%) classify pure and adulterated honey samples with different NIR spectral data.…”

Section: Introductionmentioning

confidence: 97%

Dielectric properties of honey: the potential usability for quality assessment

Pentoś

Łuczycka

2017

Eur Food Res Technol

View full text Add to dashboard Cite

Honey, as a valuable product of high price, is subject to illegal adulteration by adding water or inexpensive sweeteners as well as overheating. Therefore, cheap, quick, and reliable methods of honey quality assessment are of interest to the food industry and consumers alike. In this study, two dielectric honey parameters: relative permittivity and dielectric loss coefficient measured at frequencies below 1 MHz and at temperatures from 20 to 40 °C were considered as potential indicators of honey quality. For the data analysis, chemometric methods (artificial neural networks and two methods of extracting the contribution of independent variables) were employed. No significant differences in results of the relative contribution of input variables were found depending on frequency. Results lead to the conclusion that relative permittivity, which is significantly affected by pH, can be potentially useful for honey microbial contamination detection and dielectric loss coefficient, influenced mostly by HMF content, can be considered as an indicator of honey freshness.

show abstract

Neural networks applied to discriminate botanical origin of honeys

Cited by 80 publications

References 28 publications

Physicochemical characterization of Lavandula spp. honey with FT-Raman spectroscopy

Physicochemical characterization of Lavandula spp. honey with FT-Raman spectroscopy

Computational intelligence applied to discriminate bee pollen quality and botanical origin

Dielectric properties of honey: the potential usability for quality assessment

Contact Info

Product

Resources

About