首页 | 本学科首页   官方微博 | 高级检索  
     

真伪卷烟主流烟气指标的因子分析和聚类分析
引用本文:林 琳,曲亚玲,沈凤兰,等. 真伪卷烟主流烟气指标的因子分析和聚类分析[J]. 中国烟草学报, 2010, 16(4): 5
作者姓名:林 琳  曲亚玲  沈凤兰  
作者单位:1.北京市烟草公司质量监督检测站
摘    要:真伪卷烟由于工艺条件、原辅材料及配方的不同,烟气指标一定存在差异。本文用SM450直线型吸烟机测定了国内真、伪卷烟A的 6项主流烟气指标,并用因子分析和聚类分析对测定结果进行分析比较。因子分析的结果表明3个因子累计贡献率已达95.681%(≥85%),且总方差的58%的贡献来自第1因子,所对应真伪卷烟主流烟气成分的特征指标为一氧化碳、总粒相物和焦油。利用特征指标对20个真、伪卷烟进行层次聚类分析的结果表明当样本划分为2类时,样本的类归属情况与卷烟的真伪属性完全符合,充分说明统计分析方法在卷烟真伪鉴别工作中是一个有效的工具。 

关 键 词:卷烟   主流烟气成分分析   因子分析   烟气特征指标   聚类分析
收稿时间:2009-09-24

Factor and cluster analysis on mainstream smoke indicators in genuine and counterfeit cigarettes
Abstract:Different process conditions must lead to differences in mainstream smoke between genuine and counterfeit cigarettes. Evaluations of 6 mainstream smoke indicators were determined by SM450 linear smoking machine using both factor and cluster analyses. Result from factor analysis showed that the cumulative contribution rate of 3 factors reached 95.681%(≥85%),with the first factor account for 58% of the total variance, which means carbon monoxide, total particulate matter and tar are the characteristic indicators in mainstream smoke. Result from Hierarchical cluster analysis showed that 20 cigarettes sample including 10 counterfeit cigarettes were combined into 2 clusters which were fully consistent with the authenticity of the samples. It showed clearly that statistical analysis method is an effective tool in identifying the authenticity of the work of cigarettes. 
Keywords:
点击此处可从《中国烟草学报》浏览原始摘要信息
点击此处可从《中国烟草学报》下载免费的PDF全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号