殷勇,申晓鹏,于慧春.基于KECA+FDA的白酒电子鼻多特征鉴别方法[J].农业机械学报,2018,49(4):374-380.
YIN Yong,SHEN Xiaopeng,YU Huichun.Multi-features Identification Method of Electronic Nose Data Based on KECA+FDA for Chinese Liquors[J].Transactions of the Chinese Society for Agricultural Machinery,2018,49(4):374-380.
摘要点击次数: 2164
全文下载次数: 1024
基于KECA+FDA的白酒电子鼻多特征鉴别方法   [下载全文]
Multi-features Identification Method of Electronic Nose Data Based on KECA+FDA for Chinese Liquors   [Download Pdf][in English]
投稿时间:2017-09-24  
DOI:10.6041/j.issn.1000-1298.2018.04.044
中文关键词:  白酒  电子鼻  核熵成分分析  Fisher判别分析  多特征鉴别
基金项目:国家自然科学基金项目(31571923、31171685)
作者单位
殷勇 河南科技大学 
申晓鹏 河南科技大学 
于慧春 河南科技大学 
中文摘要:在引入基于核熵成分分析(KECA)的Fisher判别分析(FDA)方法的基础上,探究了用特征组合表征电子鼻信号时6种白酒的鉴别效果。首先,通过5种单一特征的FDA鉴别分析,筛选出积分值(INV)、相对稳态平均值(AVRS)、小波能量(WEV) 3种较优特征,然后通过它们的不同组合鉴别6种白酒,鉴别结果表明,多特征组合优于单特征,且三特征组合时的鉴别正确率最高。最后,在用INV、AVRS、WEV 3种特征值组合表征电子鼻信号的前提下,深入研究了KECA+FDA方法鉴别6种白酒的效果。当选取径向基函数(RBF)作为核函数后,采用基于矩阵最佳相似性的方法优化确定RBF核参数为16.8608时,三特征组合下测试集的鉴别正确率由FDA的79.92%提高到KECA+FDA的100%。与BP神经网络和支持向量机的鉴别结果对比,KECA+FDA方法更具优势。这说明运用KECA+FDA方法可有效提高电子鼻对6种白酒的鉴别能力。
YIN Yong  SHEN Xiaopeng  YU Huichun
Henan University of Science and Technology,Henan University of Science and Technology and Henan University of Science and Technology
Key Words:Chinese liquors  electronic nose  kernel entropy component analysis  Fisher discriminant analysis  multi-features identification
Abstract:The identification of six kinds of Chinese spirits, including similar quality using electronic nose is a complex and difficult work. In order to enhance the correct identification rate of six kinds of Chinese liquors using electronic nose (E-nose), a Fisher discriminant analysis (FDA) method based on kernel entropy component analysis (KECA) was introduced. Based on this method, the influence of different features combination representation types of E-nose signals on the discrimination result of six kinds of Chinese liquors was studied in-depth. Firstly, integral value (INV), variance (VAR), relative steady-state average value (RSAV), average differential value (ADV) and wavelet energy value (WEV) of E-nose signals were extracted as five kinds of feature values, and the FDA result of each single feature showed that the identification result based on INV, AVRS and WEV was superior to that of the other two features, respectively. Thus the features INV, AVRS and WEV were selected as subsequent analysis features. Then, for the features of INV, AVRS and WEV, when the E-nose signals were represented by random combinations based on two features or three features combination, FDA results displayed that the identification results of multi-features combinations were better than that of single feature, especially the three features combination was the best. Finally, on the premise of combining the three features to represent electronic nose signals, and the discrimination result of six kinds of Chinese liquors was deeply investigated by an introduced KECA+FDA. When the radial basis function (RBF) was selected as kernel transform function, with the help of a measuring method of matrix similarity based on Euclidean distance, the characteristic parameter of RBF was defined, which was 16.8608. And the correct identification rate of the test set samples was from 79.92% of FDA up to 100% of the KECA+FDA. Meanwhile, the discrimination result of KECA+FDA was better than that of BP neural network and support vector machine. This indicated that the KECA+FDA method can effectively improve the identification ability of the six kinds of Chinese liquors;at the same time, it also provided a feasible pattern recognition method for the identification of complex samples such as Chinese liquors by electronic nose in the future.

Transactions of the Chinese Society for Agriculture Machinery (CSAM), in charged of China Association for Science and Technology (CAST), sponsored by CSAM and Chinese Academy of Agricultural Mechanization Science(CAAMS), started publication in 1957. It is the earliest interdisciplinary journal in Chinese which combines agricultural and engineering. It always closely grasps the development direction of agriculture engineering disciplines and the published papers represent the highest academic level of agriculture engineering in China. Currently, nearly 8,000 papers have been already published. There are around 3,000 papers contributed to the journal each year, but only around 600 of them will be accepted. Transactions of CSAM focuses on a wide range of agricultural machinery, irrigation, electronics, robotics, agro-products engineering, biological energy, agricultural structures and environment and more. Subjects in Transactions of the CSAM have been embodied by many internationally well-known index systems, such as: EI Compendex, CA, CSA, etc.

   下载PDF阅读器