Tobacco Science & Technology, 2006, 11, p. 10-13, ISSN.1002-0861

Statistical analysis of routine test data of cigarette smoke

XU HAN-CHUN; LI DONG-LIANG; ZHENG JIE-QIONG
Technology Center of Xiamen Cigarette Factory, Xiamen, Fujian, China
In order to make full use of the routine test data of cigarette smoke, the data was analyzed with various statistical methods. The results indicated that: 1) the principal factors determined by principal components analysis could be used as an efficient means for weighting the items in routine cigarette smoke test; 2) the significance of difference between smokes of different batches of cigarette could be efficiently evaluated by integrated value of the items, which was derived from processing the original test data with Z fraction standardization and linear transform and combining the weight of each tested smoke item; and 3) the similarity of smoke between different cigarette brands could be evaluated with the fuzzy similitude preference method.