摘要
采用自行研制的便携式电子鼻检测系统,建立了一种能够快速辨别白酒掺假的新方法。系统检测时:首先利用传感器阵列获得白酒"指纹数据",随后通过离散小波变换(discrete wavelet transform,DWT)提取反馈信息里的特征信息,然后采用主成分分析(principal component analysis,PCA)实现对不同纯度掺假白酒样品的定性判别、采用人工蜂群优化最小二乘支持向量机(artificial bee colony least squared support vector machines,ABC-LSSVM)实现对不同纯度掺假白酒样品的定量预测。结果表明,PCA对掺假白酒区分效果较好,区分正确率高达100%; ABC-LSSVM预测模型对白酒纯度具有较高的定量预测性能,其验证集中相关系数R^2为0. 933 2,平均绝对误差MRE为6. 564 3%,均方根误差RMSE为0. 023 4。该研究可为掺假白酒的定性辨别及定量预测提供技术支持。
A new method for quickly identifying liquor adulteration was established by using a self-developed portable electronic nose detection system.When the system was detecting,the"fingerprint data"of liquor was obtained by sensor array.Then,information reflected the features was extracted by discrete wavelet transform(DWT)from the feedback.The principal component analysis(PCA)was then used to determine the quality of adulterated liquor samples with different purity.The quantitative prediction of adulterated liquor samples with different purity was realized by artificial bee colony least squared-support vector machines(ABC-LSSVM).The results showed that PCA could well-distinguish adulterated liquor.Its accuracy could be as high as 100%.ABC-LSSVM prediction model had a good quantitative prediction performance for liquor purity.Its correlation coefficient(R^2)was 0.933 2,the mean relative error(MRE)was 6.564 3%,and the root mean square error(RMSE)was 0.023 4.This study provides technical supports for qualitative discrimination and quantitative prediction of adulterated liquors.
作者
马泽亮
国婷婷
殷廷家
王志强
杨方旭
李彩虹
李钊
袁文浩
MA Zeliang;GUO Tingting;YIN Tingjia;WANG Zhiqiang;YANG Fangxu;LI Caihong;LI Zhao;YUAN Wenhao(School of Computer Science and Technology,Shandong University of Technology,Zibo 255049,China)
出处
《食品与发酵工业》
CAS
CSCD
北大核心
2019年第2期190-195,共6页
Food and Fermentation Industries
基金
国家自然科学基金项目(61473179
61701286)
山东省自然科学基金项目(ZR2018LF002
ZR2015FL003)
赛尔网络下一代互联网技术创新项目(NGII20170314)
关键词
电子鼻
白酒掺假
离散小波变换
人工蜂群优化最小二乘支持向量机
electronic nose
liquor adulteration
discrete wavelet transform
artificial bee colony-least squares support vector machine
