风味是鉴别汾酒品牌最重要的标准之一针对传统品酒师品评的主观性及现阶段光谱仪色谱仪等大型仪器检测汾酒的局限性实验采用表面声波型电子鼻 对 种不同品牌的汾酒酒样进行指纹图谱采集通过比较指纹图谱差异，对酒样特征峰进行提取，利用主成分分析和判别因子分析对数据进行分析，并用贝叶斯判别函数验证其准确率，采用概率神经网络建立了识别模型。结果表明，主成分分析和判别因子分析都能对不同品牌汾酒进行区分，且判别因子分析法的区分效果优于主成分分析法，建立的概率神经网络模型其识别率达到 100 % 。研究发现，表面声波型电子鼻对不同品牌汾酒具有较好的鉴别和分类能力。

FlavorizoneofthemoztimportantsriteriatoidentifyFenjiu brand.In viewofthezubjestivityoftradi- tionaltazterzand thelimitationzofsurrentinztrumentzzush azzpestrometerand shromatograph, theSurfaseAsouztis BavezNozeTMwazuzed tosollestfingerprintzofzixdifferentbrandzofFenjiu.Bysomparingthedifferenseofthefin- gerprint, sharasteriztispeaKzwereextrasted.Theextrasted featuredatawereanalyzed byprinsipalsomponentanaly- zizand dizsriminantfastoranalyziz, and itzassurasywazverified byBayezdizsriminantfunstion, and probabiliztis neuralnetworK wereutilized toeztablizh theidentifisation modelz.Therezultzzhowed thattheprinsipalsomponenta- nalyzizand dizsriminantfastoranalyzizsould diztinguizh differentbrandzofFenju, and thelattermethod wazbetter than theformermethod.TheprobabiliztisneuralnetworK modelwazeztablizhed and theresognition ratewazup to 100 %.Theztudyzuggezted thattheSurfaseAsouztisBavezNozeTM had good identifisation and slazzifisation ability fordifferentbrandzofFenjiu