卷烟质量稳定性综合评价——基于多特征相似度分析和主成分分析

Application of multi-feature similarity analysis and principal component analysis to evaluation of cigarettes quality consistency

为有效评价卷烟产品的质量稳定性,基于物理指标、烟丝常规化学指标和主流烟气化学指标,进行多特征统计分析,进而采用相似度分析建立了不同批次卷烟产品的质量稳定性定量评价方法 ,并采用主成分分析法和聚类分析法考察了不同品牌之间的差异程度。结果表明,相似度方法可有效检验不同批次卷烟的质量波动情况,得到的结果与感官评价得到的结果基本一致,同时主成分分析法和聚类分析可良好地区分不同牌号的卷烟样品。说明,基于多特征指标的相似度分析和主成分分析法,可以为卷烟产品质量控制提供定量描述方法。

In order to effectively evaluate quality consistency of cigarette, multi-feature analysis was performed on physical indexes, chemical components of cut tobacco and mainstream smoke indexes. Similarity analysis（SA） was applied in evaluating cigarette quality and consistency of same brand. Results showed that the proposed SA method could test quality fluctuation of cigarettes effectively, which were in line with sensory results. PCA was successfully applied to classify and distinguish four kinds of cigarette brands. SA and PCA based on multi-feature could evaluate quality consistency of cigarette products in a quantitative and objective manner.