基于Raman光谱和支持向量机回归的古井贡酒年份鉴别方法

发布时间：2018-02-16 20:40

本文关键词： Raman光谱支持向量机回归白酒年份鉴别白酒贮存时间鉴别　出处：《光谱学与光谱分析》2016年03期 　论文类型：期刊论文

【摘要】：白酒年份的快速准确鉴定是白酒品质分析的重点和难点问题之一。实现白酒年份酒的快速、准确的鉴别,对促进白酒行业的健康发展、维护消费者的合法权益具有重要意义。光谱分析法结合模式识别技术是实现白酒品质快速鉴别的首选方法之一,而Raman光谱由于其受水的影响很小且很少或不需要样品前处理,在白酒分析中具有广阔的发展空间。因此,采用Raman光谱和支持向量机回归(SVR)建立数据分析模型,用于不同年份白酒的年份鉴定和同一年份不同贮存时间的白酒年份鉴定。该研究创新之处主要包括如下三个方面:(1)应用Raman光谱对白酒品质进行分析,在分析方法的应用上具有一定的创新之处。(2)研究白酒的年份鉴定问题,在研究对象的选择上,具有一定的创新之处。(3)建立基于回归框架的白酒年份与年份指数对应关系,实现白酒年份识别及预测,不仅可以有效鉴别白酒年份,同时可用于鉴别白酒贮存时间,因此,在分析方法的确定和应用上,具有一定的创新之处。实验中采用古井贡5年、8年、16年及26年系列年份酒进行了实证分析,数据分析实验结果表明,所建立的基于Raman光谱和SVR的白酒年份鉴别分析流程和方法,对鉴别不同年份的白酒,以及同一年份不同贮存时间的白酒样品(包括对数据库内已有样本年份的鉴别,以及对数据库内没有的盲样的年份预测),均取得较好的应用效果,相比于其他常用回归分析方法具有明显的优越性,可以为白酒年份酒分析提供一定的技术支持。
[Abstract]:The rapid and accurate identification of liquor year is one of the key and difficult problems in liquor quality analysis. To realize the rapid and accurate identification of liquor year wine will promote the healthy development of liquor industry. It is of great significance to protect the legitimate rights and interests of consumers. Spectral analysis combined with pattern recognition is one of the preferred methods for rapid identification of liquor quality, while Raman spectrum is less affected by water and requires little or no sample pretreatment. Therefore, Raman spectrum and support vector machine (SVM) were used to establish the data analysis model. It is used to identify the year of liquor in different years and the different storage time in the same year. The main innovation of this research includes the following three aspects: 1) the quality of liquor was analyzed by Raman spectrum. In the application of analytical method, there are some innovations in the application of this method. (2) to study the year identification problem of liquor, and to establish the corresponding relationship between the year and the year index based on regression framework in the selection of the object of study. To realize the identification and prediction of liquor years, not only the liquor years can be effectively identified, but also the storage time of liquor can be identified. Therefore, in the determination and application of analytical methods, In the experiment, the series of vintage wines of Gujinggong for 5, 8, 16 and 26 years were used for empirical analysis. The results of data analysis showed that the process and method of year identification and analysis based on Raman spectrum and SVR were established. Good results were obtained in identifying liquor samples of different years and different storage times in the same year (including the years of existing samples in the database and predicting the years of blind samples not in the database). Compared with other commonly used regression analysis methods, it has obvious advantages and can provide some technical support for liquor analysis.
【作者单位】：南京财经大学管理科学与工程学院;江苏省质量安全工程研究院;
【基金】：国家重大科学仪器开发专项(2013YQ090703) 国家自然科学基金重点项目(71433006);国家自然科学基金面上项目(61373058,71373117) 国家质量监督检验检疫总局公益性行业科研专项(2012104009)资助
【分类号】：O657.37;TS262.3

【相似文献】