线性判别分析法在塑料食品包装容器溶出物的风险聚类分析中的应用  

Risk-clustering on Residues from Plastic Food Package by Linear Discriminant Analysis

作　　者：林雨霏[1] 王岩[1] 纪雷[1] 孙健[1] 刘心同[1] 崔鹤[1] 宋振乾[1] 

机构地区：[1]山东出入境检验检疫局,青岛266002

出　　处：《包装工程》2009年第5期6-8,共3页Packaging Engineering

基　　金：山东省科技攻关项目(2008GG10008005);国家质检总局重大科研项目基金(2008IK256)资助

摘　　要：报道了为考察塑料食品包装容器蒸发残渣因不同材料造成的质量差异,并对不同材质蒸发残渣指标进行风险聚类,提高检测针对性的方法。对塑料食品包装容器在水、乙酸、乙醇、正己烷浸泡条件下的蒸发残渣含量进行测定,使用四分位稳健统计描述来表征不同使用条件下的溶出风险,使用线性判别法分析对不同材质塑料包装容器蒸发残渣进行了风险聚类分析。结果表明:包装容器的油性溶出风险最高,水的溶出风险最低;包装容器蒸发残渣含量水平可基本划为两类,即"PET-PP-PE"类与"PS-其它类";采用适合的数据描述及多元统计分析方法,可有效地加强食品包装风险监控力度,提高监控效率。Contents of residues from plastic food package under different immersed conditions, i. e. water, acetic acid, ethanol and n-hexane, were investigated by multivariate statistical techniques respectively. The discrepancy between conventional data description and quartile robust description were discussed. Risk-clustering of plastic for food package based on material/component classification was investigated further by linear discriminant analysis （LDA）. Results showed that the residue levels in all plastic food packages can be divided into two categories, i.e. PET-PP-PE, PS and others; Application of suitable data description and multivariate statistical methods can effective improve the strength and efficiency of risk control of food packaging.

关 键 词：塑料食品包装 蒸发残渣 四分位稳健统计描述 线性判别分析 

分 类 号：TS255.3[轻工技术与工程—农产品加工及贮藏工程] TB484.3[轻工技术与工程—食品科学与工程]