基于神经网络预测啤酒老化过程中感官得分的初步研究

Prediction of sensory evaluation scores in beer aging process based on artificial neural network

选取8种不同种类新鲜啤酒通过实验手段模拟其老化过程,对老化后的啤酒进行感官品评并分析了新鲜啤酒的老化相关指标。使用灰色关联分析法分析了啤酒的感官评价得分和啤酒老化相关指标之间的关联度,并根据其影响程度的大小进行了排序。把新鲜啤酒的感官得分和啤酒老化相关指标作为输入数据,不同老化时间的感官得分作为输出数据,使用BP神经网络建立了模型啤酒老化过程中感官得分模型并进行预测。感官得分最小相对误差为1.51%。神经网络方法能够有效地预测啤酒老化过程中的感官变化。

The correlation between the aging-related index and sensory evaluation of fresh beer was analyzed by means of simulation experiments of their aging process.Using the evaluation scores of fresh beer sensory and the index of fresh beer aging-related as input and different time aging beer sensory evaluation scores as output,and the back-propagation neural network was used to establish the prediction model.The relative error of predictive model is 1.51%,indicating that the model is adapted to forecast beer sensory evaluation changes in the process of beer aging.