摘要
为了提高6种食醋的电子鼻鉴别能力,该文提出了一种基于多变量分析的食醋电子鼻信号多特征表征策略。初选不同的特征表征电子鼻信号,构建电子鼻信号的初始特征矩阵。采取载荷分析进行电子鼻传感器阵列优化,优选了12个气敏传感器的响应数据进行后续分析。为消除各传感器响应信号之间的相关性,对优选阵列的特征矩阵进行主成分分析(principal component analysis,PCA),并利用WilksΛ统计量选择鉴别能力最优的主成分子阵。在选择最优主成分子阵的基础上,以生成主成分的每一个原始特征变量为对象,计算每一个原始特征变量在主成分子阵中的贡献系数绝对值之和,且根据系数绝对值之和从大到小排序;同时,根据不同和值的指定,形成了不同容量的原始特征变量集。最后,借助于Fisher判别分析(Fisher discriminant analysis,FDA)探索了不同容量原始特征变量集的鉴别结果,确定了最佳的原始特征变量集。结果表明,特征选择前后传感器信号的表征特征发生了明显变化,最终采用48个特征参量实现了对食醋电子鼻信号的有效表征。在48个特征参量表征条件下,同时运用FDA和BP神经网络(back propagation neural network,BPNN)对6种食醋进行了鉴别分析,训练集的鉴别正确率分别在93%和98%以上,测试集的鉴别正确率也分别达到了90%和93%以上。另外,利用巴氏距离进一步揭示了样品间的可分离程度及FDA与BPNN结果的可信性。研究结果可为电子鼻信号多特征表征提供了一种新思路。
In order to enhance discrimination ability of electronic nose(E-nose) for six kinds of vinegars, a multi-features representation strategy for E-nose data of vinegar samples based on multivariable analysis is proposed in this paper. Firstly, initial feature matrix, which was composed of six kinds of features extracted from E-nose data, was dealt with loadings analysis so as to optimize gas sensors, and then kept 12 gas sensors for next analysis. For eliminating correlation between response signals of gas sensors, feature matrix of 12 sensors array was carried out with principal component analysis(PCA), and generated principal component(PC) variables(PC variable(s) for short) for constructing Wilks Λ-statistic. Subsequently, Wilks Λ value of each PC variable was obtained. As we all known, the smaller the value of Λ, the higher separation ability of the calculated PC variables; in other words, some PC variables corresponding to larger Λ values should be eliminated due to their lower separation ability. Generally speaking, Wilks Λ-statistic was adopted to get principal component sub-matrix that was beneficial to identification of vinegar samples. On the basis of obtaining principal component sub-matrix, considering that each PC variable was a linear combination of all original feature variables, as for each original feature variable, the contribution quantity of original feature variable to all obtained PC variables may be as choosing criterion. So taking each original feature variable as an object, and the sum of absolute values of combination coefficients corresponding to each original feature variables would be calculated according to obtained principal component sub-matrix, and the sums corresponding to different original feature variables were sorted from large to small, and the greater the sum, the higher possibility for the corresponding original feature variables to be chosen. Meanwhile, according to different designation values for the sum of coefficient absolute values of each original feature variable to all picked PC variables, different original feature variable sets could be formed. With the help of Fisher discriminant analysis(FDA), after correct discrimination rates of different original feature variable sets were calculated and compared, optimal original feature variables set was determined. The results showed that representation feature variables for gas sensors were extremely different from initial ones. In view of the proposed feature selection strategy, 48 features were selected to characterize E-nose signals of vinegar samples at final. In order to verify and explain the application effect of feature selection strategy and the rationality of selected 48 characteristic parameters for vinegar samples, FDA and back propagation neural network(BPNN) were employed to discriminate six kinds of vinegar samples, and correct discrimination rates of FDA and BPNN were over 93% and 98% in training sets, respectively; corresponding test sets were also over 90% and 93%, respectively. In addition, Bhattacharyya distance was also employed further to explain the separability between six kinds of vinegar samples and illustrate the reliability of FDA and BPNN results. As a result, the proposed feature selection strategy is effective and feasible, which provides a new idea for multi-features representation of E-nose data.
引文
[1]阳飞,张华山.食醋及其营养保健功能研究进展[J].中国调味品,2017,42(5):171-175.Yang fei,Zhang Huashan.Research progress on nutrition and health care function of vinegar[J].China Condiment,2017,42(5):171-175.(in Chinese with English abstract)
[2]梅辉.食醋的营养以及保健功能研究[J].现代食品,2017,11(22):7-8.Mei Hui.Study on nutrition and health function of vinegar[J].Modern Food,2017,11(22):7-8.(in Chinese with English abstract)
[3]Giudici P,陈福生,杨浩然,等.中国谷物醋的感官风味特征分析[J].中国酿造,2017,36(9):1-5.Giudici P,Chen Fusheng,Yang Haoran,et al.Analysis of sensory flavor characteristics of Chinese cereal vinegar[J].China Brewing,2017,36(9):1-5.(in Chinese with English abstract)
[4]聂志强,韩玥,郑宇,等.宏基因组学技术分析传统食醋发酵过程微生物多样性[J].食品科学,2013,34(15):198-203.Nie Zhiqiang,Han Yue,Zheng Yu,et al.Metagenomic analysis of microbial diversity in the traditional vinegar fermentation process[J].Food Science,2013,34(15):198-203.(in Chinese with English abstract)
[5]Li S,Li P,Feng F,et al.Microbial diversity and their roles in the vinegar fermentation process[J].Applied Microbiology&Biotechnology,2015,99(12):4997-5024.
[6]李华北,陈斌,赵杰文,等.用小波变换技术提高食醋近红外光谱分析的精度[J].农业工程学报,2000,16(6):114-117.Li Huabei,Chen Bin,Zhao Jiewen,et al.Application of wavelet transform technology to the improvement of analyzing accuracy of vinegar near-infrared spectrum[J].Transactions of the Chinese Society of Agricultural Engineering(Transactions of the CSAE),2000,16(6):114-117.(in Chinese with English abstract)
[7]李河,林勤保,田海娇.GC-MS同时测定食醋中的16种增塑剂和7种抗氧化剂[J].食品科学,2013,34(16):143-148.Li He,Lin Qinbao,Tian Haijiao.Simultaneous determination of 16 plasticizers and 7 antioxidants in vinegar by GC-MS[J].Food Science,2013,34(16):143-148.(in Chinese with English abstract)
[8]Men H,Liu H Y,Wang L,et al.Optimization of electronic nose sensor array and its application in the classification of vinegar[J].Sensor World,2010,121-122:27-32.
[9]Yong Y,Yu H,Bing C,et al.A sensor array optimization method of electronic nose based on elimination transform of Wilks statistic for discrimination of three kinds of vinegars[J].Journal of Food Engineering,2014,127(4):43-48.
[10]张厚博,梅笑冬,赵万,等.用于食醋品质预评价的电子鼻研究[J].传感器与微系统,2013,32(3):62-64.Zhang Houbo,Mei Xiaodong,Zhao Wan,et al.Research on electronic nose for pre-evaluation of vinegar quality[J].Transducer and Microsystem Technologies,2013,32(3):62-64.(in Chinese with English abstract)
[11]张顺平,张覃轶,李登峰,等.电子鼻技术在食醋识别中的应用[J].传感技术学报,2006,19(1):104-107.Zhang Shunping,Zhang Qinyi,Li Dengfeng,et al.Research on vinegars identification by electronic nose[J].Chinese journal of sensors and actuators,2006,19(1):104-107.(in Chinese with English abstract)
[12]徐赛,陆华忠,周志艳,等.基于电子鼻的果园荔枝成熟阶段监测[J].农业工程学报,2015,31(18):240-246.Xu Sai,Lu Huazhong,Zhou Zhiyan,et al.Electronic nose monitoring mature stage of litchi in orchard[J].Transactions of the Chinese Society of Agricultural Engineering(Transactions of the CSAE),2015,31(18):240-246.(in Chinese with English abstract)
[13]于慧春,褚冰,殷勇.食醋电子鼻检测中一种特征参量评价方法[J].农业工程学报,2013,29(3):258-264.Yu Huichun,Chu Bing,Yin Yong.Evaluation method of feature vector in vinegar identification by electronic nose[J].Transactions of the Chinese Society of Agricultural Engineering(Transactions of the CSAE),2013,29(3):258-264.(in Chinese with English abstract)
[14]尹芳缘,黄洁,王敏敏.用电子鼻区分霉变燕麦及其传感器阵列优化[J].农业工程学报,2013,29(20):263-269.Yu Fangyuan,Huang Jie,Wang Minmin,et al.Discrimination of mildewed oats using electronic nose and optimization of its sensor array[J].Transactions of the Chinese Society of Agricultural Engineering(Transactions of the CSAE),2013,29(20):263-269.(in Chinese with English abstract)
[15]Yin Y,Hao Y,Yu H,et al.Detection potential of multi-features representation of E-Nose data in classification of moldy maize samples[J].Food&Bioprocess Technology,2017,10(12):1-14.
[16]邵小龙,张蓝月,宋伟,等.籼稻储藏品质的电子鼻快速检测方法研究[J].中国粮油学报,2014,29(4):104-107.Shao Xiaolong,Zhang Lanyue,Song Wei,et al.Rapid detection method for stored indica rice by electronic nose[J].Journal of the Chinese Cereals and Oils Association,2014,29(4):104-107.(in Chinese with English abstract)
[17]Xu K,Wang J,Wei Z,et al.An optimization of the MOSelectronic nose sensor array for the detection of Chinese pecan quality[J].Journal of Food Engineering,2017,203:25-31.
[18]何金鑫,郜海燕,穆宏磊,等.山核桃氧化过程中品质指标变化的电子鼻快速检测[J].农业工程学报,2017,33(14):284-291.He Jinxin,Gao Haiyan,Mu Honglei,et al.Rapid detection of quality parameters change in hickory oxidation process by electronic nose[J].Transactions of the Chinese Society of Agricultural Engineering(Transactions of the CSAE),2017,33(14):284-291.(in Chinese with English abstract)
[19]傅润泽,沈建,王锡昌,等.基于神经网络及电子鼻的虾夷扇贝鲜活品质评价及传感器的筛选[J].农业工程学报,2016,32(6):268-275.Fu Runze,Shen Jian,Wang Xichang,et al.Quality evaluation of live Yesso scallop and sensor selection based on artificial neural network and electronic nose[J].Transactions of the Chinese Society of Agricultural Engineering(Transactions of the CSAE),2016,32(6):268-275.(in Chinese with English abstract)
[20]徐克明,王俊,邓凡霏,等.用于山核桃陈化时间检测的电子鼻传感器阵列优化[J].农业工程学报,2017,33(3):281-287.Xu Keming,Wang Jun,Deng Fanfei,et al.Optimization of sensor array of electronic nose for aging time detection of pecan[J].Transactions of the Chinese Society of Agricultural Engineering(Transactions of the CSAE),2017,33(3):281-287.(in Chinese with English abstract)
[21]殷勇,郝银凤,于慧春.基于多特征融合的电子鼻鉴别玉米霉变程度[J].农业工程学报,2016,32(12):254-260.Yin Yong,Hao Yinfeng,Yu Huichun.Identification method for different moldy degrees of maize using electronic nose coupled with multi-features fusion[J].Transactions of the Chinese Society of Agricultural Engineering(Transactions of the CSAE),2016,32(12):254-260.(in Chinese with English abstract)
[22]殷勇,吴文凯,于慧春.独立分量分析融合小波能量阈值的电子鼻信号去漂移方法[J].农业工程学报,2014,30(24):325-331.Yin Yong,Wu Wenkai,Yu Huichun.Drift elimination method of electronic nose signals based on independent component analysis coupled with wavelet energy threshold value[J].Transactions of the Chinese Society of Agricultural Engineering(Transactions of the CSAE),2014,30(24):325-331.(in Chinese with English abstract)
[23]孙香丽,殷勇.乳制品电子鼻分类中传感器阵列的一种优化方法[J].传感技术学报,2008,21(7):1124-1127.Sun Xiangli,Yin Yong.Optimization method of sensor array for dairy products identification by electronic nose[J].Chinese journal of sensors and actuators,2008,21(7):1124-1127.(in Chinese with English abstract)
[24]Savitzky A,Golay M J E.Smoothing and differentiation of data by simplified least squares procedures[J].Analytical Chemistry,1964,36(8):1627-1639.
[25]Zhang S,Xia X,Xie C,et al.A method of feature extraction on recovery curves for fast recognition application with metal oxide gas sensor array[J].IEEE Sensors Journal,2009,9(12):1705-1710.
[26]徐赛,周志艳,罗锡文.常规稻与杂交稻谷的仿生电子鼻分类识别[J].农业工程学报,2014,30(9):133-139.Xu Sai,Zhou Zhiyan,Luo Xiwen.Classification and recognition of hybrid and inbred rough rice based on bionic electronic nose[J].Transactions of the Chinese Society of Agricultural Engineering(Transactions of the CSAE),2014,30(9):133-139.(in Chinese with English abstract)
[27]张婷婷,孙群,杨磊,等.基于电子鼻传感器阵列优化的甜玉米种子活力检测[J].农业工程学报,2017,33(21):275-281.Zhang Tingting,Sun Qun,Yang Lei,et al Vigor detection of sweet corn seeds by optimal sensor array based on electronic nose[J].Transactions of the Chinese Society of Agricultural Engineering(Transactions of the CSAE),2017,33(21):275-281.(in Chinese with English abstract)
[28]程绍明,王俊,王永维,等.基于电子鼻技术的不同特征参数对番茄苗早疫病病害区分效果影响的研究[J].传感技术学报,2014,27(1):1-5.Cheng Shaoming,Wang Jun,Wang Yongwei,et al.Research on distinguishing tomato seedling infected with early blight disease using different characteristic parameters by electronic nose[J].Chinese Journal of Sensors and Actuators,2014,27(1):1-5.(in Chinese with English abstract)
[29]Bekker A,Roux J J J,Arashi M.Exact nonnull distribution of Wilks'statistic:The ratio and product of independent components[J].Journal of Multivariate Analysis,2011,102(3):619-628.
[30]Peng X,Zhang L,Tian F,et al.A novel sensor feature extraction based on kernel entropy component analysis for discrimination of indoor air contaminants[J].Sensors&Actuators A Physical,2015,234(11):143-149.
[31]高惠璇.应用多元统计分析[M].北京:北京大学出版社,2005.
[32]张连蓬,李行,陶秋香.高光谱遥感影像特征提取与分类[M].北京:测绘出版社,2012..