Application of Principal Component Analysis and Cluster Analysis in Tobacco Quality Evaluation

Combining with appearance quality and chemical analysis, the quality of 45 samples from Xinyang tobacco growing area in Henan was evaluated by principal component analysis and cluster analysis.The results showed that the appearance quality of samples was fine in general, while they were slightly weak in oil, colority, chromatism and leaf surface uniformity.The contents of the majority of chemical components were up to the standard of quality tobacco leaf, though total nitrogen and K were lower and starch was slightly higher.The samples were divided into 3 groups by cluster analysis, which integrated 8 appearance indexes with 8 chemical indexes and rejected repeated information among the 16 indexes.