A Research on Mathematical Model and Prediction Method for Raw Tobacco Material Quality Based on Cluster Classification

Abstract The chemical components of 189 kinds of Chinese tobacco materials were analyzed. There were obvious correlations between the total sugar content and reducing sugar content, total sugar content and total nitrogen content, reducing sugar content and nitrogen content, nicotine content and nitrogen content, respectively. There were certain correlations between chemical components and quality evaluation score of the tobacco materials. Total sugar and reducing sugar content showed similar correlations, while nitrogen, chloride ion and nicotine content showed similar correlations. The correlation between each single component and the quality evaluation score was not strong, and it was difficult to be used as a classification standard. Using the cluster classification method, the tobacco materials were well classified into three clusters. For each cluster, a reasonable relationship between the quality evaluation score and the chemical component contents of the raw tobacco material was set up. The method for rapid evaluation of tobacco quality score were established based on these results.