Artificial neural networks (ANN): prediction of sensory measurements from instrumental data

Carvalho, Naiara Barbosa; Minim, Valéria Paula Rodrigues; Silva, Rita de Cássia dos Santos Navarro; Lúcia, Suzana Maria Della; Minim, Luís Antônio

doi:10.1590/s0101-20612013000400018

Cited by 23 publications

(18 citation statements)

References 36 publications

(38 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Therefore, an artificial neural network was designed in this study, taking into account the type of carrier, its concentration, total solids content of encapsulant solution, and its density and viscosity as the input parameters and all the experimentally determined results as the output parameters. The most suitable network was selected based on training, test and validation preferences and errors, as previously stated by Carvalho et al (2013) and Di Scala et al (2013), and had a hidden activation functions identity and logistic output, with the maximum test error of 0.027 achieved in the test period, and was selected for analysis of the spray-drying process . Global sensitivity coefficient, as an indicator of influence of the particular input parameter on the output variable, is defined as the ratio of variances of individual parameter relative to the total variance.…”

Section: Artificial Neural Networkmentioning

confidence: 99%

Efficiency Assessment of Natural Biopolymers as Encapsulants of Green Tea (Camellia sinensis L.) Bioactive Compounds by Spray Drying

Belščak‐Cvitanović

Lević

Kalušević

et al. 2015

Food Bioprocess Technol

View full text Add to dashboard Cite

In this study a bottom-up approach of designing functional ingredients from green tea extract is proposed by a systematic investigation of 12 different natural biopolymers and their efficiency as carrier materials of green tea bioactive compounds by spray drying at low temperature (130°C). Screening of carriers revealed that inulin and whey proteins provide the highest product yields (67.04 and 65.18 %, respectively) and, accompanied with pectin, also the highest total polyphenols (67.5-82.2 %) and flavan-3-ols (93.7-75.9 %) loading capacity. Up to 162 mg/g of (−)−epigallocatechin gallate (EGCG) was achieved, while low-caffeine contents (<5 mg/g) indicated the potential of obtaining lowcaffeine functional ingredients. Employing alginate, carageenan and gums (acacia gum and xanthan) enabled the best colour preservation and highest chlorophyll content. Reconstituted green tea microencapsulates comprising modified starch, inulin or carageenan exhibited the lowest bitterness and astringency and the highest green tea flavour intensity as the most favourable sensory properties. An artificial neural network (ANN) designed based on the experimentally obtained results revealed hydrocolloid gums as the best encapsulants for achieving good physical properties, high EGCG contents and prolonged dissolution/release profiles while pectin, inulin and modified starch as the optimal ones in terms of the product yields, loading capacities and sensory properties. This indicates that a formulation comprising a combination of all of those biopolymers would provide potentially functional ingredients with encapsulated green tea phytochemicals, retained colour and improved sensory properties.

show abstract

Section: Artificial Neural Networkmentioning

confidence: 99%

Efficiency Assessment of Natural Biopolymers as Encapsulants of Green Tea (Camellia sinensis L.) Bioactive Compounds by Spray Drying

Belščak‐Cvitanović

Lević

Kalušević

et al. 2015

Food Bioprocess Technol

View full text Add to dashboard Cite

show abstract

“…A topologia ou arquitetura de uma rede neural consiste na organização e disposição dos neurônios, formando camadas ou grupos de neurônios mais ou menos distantes da entrada e saída da rede [42]. Nesse sentido, os parâmetros fundamentais da rede são: o número de camadas, o número de neurônios por camada, o grau de conectividade e o tipo de conexões entre os neurônios.…”

Section: Topologiaunclassified

“…Entretanto, em um grande número dessas redes existe também a possibilidade de conectar a saída dos neurônios das camadas subsequentes à entrada das camadas anteriores; essas conexões são chamadas de conexões ou feedback de retorno [42].…”

Section: Topologiaunclassified

“…Nas redes neurais artificiais (inspiradas na biologia), a função de ativação é geralmente uma abstração que representa uma taxa de ativação do potencial de ativação na célula. Em sua forma simplificada, essa função é binária, ou seja, o neurônio é ativado ou não [42]. Nesse caso, o vetor de entrada é transformado para refletir um limite de decisão de hifélica com base em algumas poucas entradas de treinamento chamadas vetores de suporte [41].…”

Section: Função De Ativaçãounclassified

“…Normalmente, o treinamento para, quando o cálculo do erro quadrado em todos os exemplos de treinamento atingiu um mínimo ou quando, para cada um dos exemplos dados, o erro observado está abaixo de um certo limite. Para observar este processo, a maioria das ferramentas de redes neurais exibem estes erros usando gráficos [42].…”

Section: Treinamentounclassified

See 2 more Smart Citations

Modelagem De Corpos Estranhos Magnéticos E Solução Do Problema Inverso Por Redes Neurais

HUACASI¹

View full text Add to dashboard Cite

show abstract

Multiple regression models and Artificial Neural Network (ANN) as prediction tools of changes in overall quality during the storage of spreadable processed Gouda cheese

Stangierski

Weiss

Kaczmarek

2019

Eur Food Res Technol

View full text Add to dashboard Cite

The aim of the study was to compare the ability of multiple linear regression (MLR) and Artificial Neural Network (ANN) to predict the overall quality of spreadable Gouda cheese during storage at 8 °C, 20 °C and 30 °C. The ANN used five factors selected by Principal Component Analysis, which was used as input data for the ANN calculation. The datasets were divided into three subsets: a training set, a validation set, and a test set. The multiple regression models were highly significant with high determination coefficients: R 2 = 0.99, 0.87 and 0.87 for 8, 20 and 30 °C, respectively, which made them a useful tool to predict quality deterioration. Simultaneously, the artificial neural networks models with determination coefficient of R 2 = 0.99, 0.96 and 0.96 for 8, 20 and 30 °C, respectively were built. The models based on ANNs with higher values of determination coefficients and lower RMSE values proved to be more accurate. The best fit of the model to the experimental data was found for processed cheese stored at 8 °C.

show abstract

Artificial neural networks (ANN): prediction of sensory measurements from instrumental data

Cited by 23 publications

References 36 publications

Efficiency Assessment of Natural Biopolymers as Encapsulants of Green Tea (Camellia sinensis L.) Bioactive Compounds by Spray Drying

Efficiency Assessment of Natural Biopolymers as Encapsulants of Green Tea (Camellia sinensis L.) Bioactive Compounds by Spray Drying

Modelagem De Corpos Estranhos Magnéticos E Solução Do Problema Inverso Por Redes Neurais

Multiple regression models and Artificial Neural Network (ANN) as prediction tools of changes in overall quality during the storage of spreadable processed Gouda cheese

Contact Info

Product

Resources

About