BP网络与RBF网络在感官评估中应用比较

Comparison with BP and RBF neural network used in sensory evaluation

介绍BP网络与RBF网络的方法原理以及二者在感官评估中的应用。利用企业提供的单料烟感官数据设计建立BP网络和RBF网络模型,并利用模型预测单料烟感官质量指标,然后通过行业专家提供的符合度公式对建立的模型进行评估,评估结果以百分比的形式展示。实验结果表明:在感官评估应用中,RBF网络模型预测性能优于BP网络模型。

The theories of BP and RBP network were introduced,meanwhile their applications in sensory evaluation were discussed.The model of BP and RBF neural network were desigined about tobacco data provided by industry,then the model was used to predict the sensory index about tobacco.Finally,the models of BP and RBF networks were evaluated using the fitting level formula provided by industry experts and the result was displayed in the form of percentage.The final result reveals that RBF model is more excellent than BP model in the application of sensory evaluation.