Artificial neural network: a powerful tool in associating phenolic compounds with antioxidant activity of grape juices

In vitro techniques are essential to assess the antioxidant potential of foods, although methods with different action mechanisms make troublesome data analysis. This article describes use artificial neural network (ANN) associate phenolic compounds activity in (AOX) grape juices. A multilayer perceptron (MLP) ANN was obtained 28 phenolics quantified, as input layers, and AOX measuring by DPPH, ABTS, FRAP, H2O2, β-carotene/linoleic acid bleaching assay (βCLA) methods, output layers. To improve discussion food sciences, results were compared Pearson’s correlation principal component analysis (PCA), largely used studies. technique showed correlations between some compounds, but limitations. PCA proved be a more powerful method than correlation, it positively associated 13 four out five methods. The MLP-ANN allowed simultaneous association 19 individual phenolics, while single hidden layer predicted 15 all power was: > Pearson. It evidenced that is tool for screening antioxidants systems, which applicable health interests.