基于径向基函数神经网络探索风味物质对啤酒风味的预测

Prediction of beer flavor with flavor compounds by radial basis function neural network model

以啤酒中对风味贡献较大的高级醇、酯、有机酸、酒精、苦味质共13种风味物质作为输入指标,以感官品评评分为输出指标,利用径向基函数(Radial basis function,RBF)神经网络(neural network)对国内外20种啤酒的风味进行预测,准确率达75%,与其他相同样本数的神经网络相比较,准确率有较大提高。说明啤酒中这13种风味物质与综合风味之间存在密切关系,以利于啤酒生产厂家对啤酒整体风味的控制。

Using 13 flavor compounds including higher alcohol, ester, organic acid, ethanol and bitter principle as input and sense evaluation scores as output, the flavor of 20 beers made inside or outside of China was predicted by radial basis function neural network. The predictive accuracy was 75%, which was much higher than the other prediction by neural network with the same sample scale. It was indicated that there was close relationship existed among the 13 flavor compounds and beer flavor.