Predicting partition coefficients of migrants in food simulant/polymer systems using adaptive neuro-fuzzy inference system

Abstract

A contaminação de alimentos pela migração de aditivos de baixo peso molecular em alimentos processados industrialmente pode ser resultado do contato direto entre a embalagem e o alimento. A concentração do aditivo que migra do material da embalagem para o alimento está relacionada com as propriedades estruturais do aditivo, bem como com a natureza do material empregado na embalagem. O objetivo deste estudo é desenvolver um modelo QSPR pela adaptação do sistema de interferência neuro-fuzzy (ANFIS) a fim de predizer o valor do coeficiente de partição K, no sistema de estudo, embalagem/alimento. Para tal, foram investigados 44 coeficientes de partição em vários sistemas, assim constituídos: 4 de simuladores alimentares, 6 de migrantes alimentares e 2 de embalagens. Um conjunto de 6 descritores moleculares, representando várias características dos simuladores de alimentos (2 descritores), dos migrantes (3 descritores) e de polímeros (1 descritor) foi empregado como a série de dados para avaliar esse estudo. Esta série de dados foi dividida em três subconjuntos: treinamento, teste e predição. A técnica de modelagem ANFIS foi aplicada pela primeira vez neste campo de estudos relacionado com alimento/embalagem. O resultado desta modelagem forneceu um RMSE de 0,0006 e o coeficiente de correlação (R 2 ) para o ensaio da predição foi de 0,9920.Food contaminations by migration of low molecular weight additives into foodstuffs can result from direct contact between packaging materials and food. The amount of migration is related to the structural properties of the additive as well as to the nature of packaging material. The goal of this study is to develop a quantitative structure-property relationship (QSPR) model by the adaptive neuro-fuzzy inference system (ANFIS) for prediction of the partition coefficient K in food/ packaging system. The partition coefficients of a set of 44 systems consisted of 4 food simulants, 6 migrants and 2 packaging materials were investigated. A set of 6 molecular descriptors representing various structural characteristics of food simulants (2 descriptors), migrants (3 descriptors) and polymers (1 descriptor) was used as data set. This data set was divided into three subsets: training, test and prediction. ANFIS as a new modeling technique was applied for the first time in this field. The final model has a root mean square error (RMSE) of 0.0006 and correlation coefficient (R 2 ) for the prediction set of 0.9920.

Quantitative structure-property relationships of electroluminescent materials: Artificial neural networks and support vector machines to predict electroluminescence of organic molecules

“…The best support vector machine model (R 2 = 0•90167, MAE = 0•00117, RMSE = 0•03420 and q 2 = 0•85885) exhibit better performance than the best artificial neural networks model (R 2 = 0•80395, MAE = 0•00395, RMSE = 0•06285 and q 2 = 0•80311). As the higher values of R 2 and q 2 indicate higher predictive power of the model (Shahbazikhah et al 2011), it is clear that the support vector machine model has higher predictive power than artificial neural networks model. Support vector machine regression is suitable to be applied in QSPR studies with small datasets, since it can interpret the nonlinear relationships between a molecular structure and its properties (Juna et al 2010), while artificial neural networks generally need larger datasets to exhibit good performance.…”

Molecular docking, molecular dynamics simulation, and QSAR model on potent thiazolidine-4-carboxylic acid inhibitors of influenza neuraminidase

“…This procedure was carried out 200 times. The theory behind these validation methods has been sufficiently described elsewhere (Jalali-Heravi et al, 2009;Shahbazikhah et al, 2011). The statistical results are given in Table 3 shows the stepwise MLR-LS-SVR and CART-LS-SVR models possessed all the criteria to be considered as a predictive model.…”

Quantitative structure–activity relationship analysis of human neutrophil elastase inhibitors using shuffling classification and regression trees and adaptive neuro-fuzzy inference systems

“…It is based on the 'fuzzy partition' of input space and it can be viewed as the expansion of a piecewise linear partition [30]. A detailed description of the theory behind an ANFIS has been adequately described elsewhere [29][30][31].…”