摘要
通过对豆腐乳理化指标和感官特征进行分析,利用感官评判的方式提取豆腐乳的前4个感官主成分和前4个理化主成分,对豆腐乳感官品质影响较大的前4个主成分贡献率分别为49.66%、16.33%、12.85%、6.49%;对豆腐乳理化特性影响较大的前4个主成分贡献率分别为32.67%、21.33%、17.85%、7.22%。以主成分分析结果为输入参数,利用遗传算法建立神经网络预测模型,对豆腐乳的综合评价进行预测,与感官评分和理化特性的主成分综合评分结果进行对比,数据表明,二者之间的偏差值最大为9.92%,具有较大的预测可信度,表明利用遗传算法能够对豆腐乳的感官综合评价进行有效预测。
Based on the analysis of physicochemical indexes and sensory characteristics of fermented bean curd,the first four sensory principal components and the first four physicochemical principal components of fermented bean curd are extracted by sensory evaluation method.The contribution rate of the first four principal components which have great influence on the sensory quality of fermented bean curd is 49.66%,16.33%,12.85%,6.49%respectively;the contribution rate of the first four principal components which have great influence on the physicochemical properties of fermented bean curd is 32.67%,21.33%,17.85%,7.22%respectively.Taking the results of principal component analysis as the input parameters,a neural network prediction model is established by using genetic algorithm to predict the comprehensive evaluation of fermented bean curd.Compared with the comprehensive evaluation results of principal components of sensory evaluation scores and physicochemical properties,the data show that the maximum deviation value between the two is 9.92%,which has a high prediction reliability.The results show that the genetic algorithm could effectively predict the sensory comprehensive evaluation of fermented bean curd.
作者
李晓敏
董素芬
LI Xiao-min;DONG Su-fen(Baoding Vocational and Technical College,Baoding 071051,China;Hebei Agricultural University,Baoding 071051,China)
出处
《中国调味品》
CAS
北大核心
2021年第9期151-153,160,共4页
China Condiment
基金
河北省教育厅课题(C20190336)。
关键词
豆腐乳
遗传算法
感官评价
主成分分析
理化特性
fermented bean curd
genetic algorithm
sensory evaluation
principal component analysis
physicochemical properties
