Abstract
Based on ideas and methods of organoleptic evaluation on agricultural commodities, the article establishes the quantitative indicators that can make effect evaluation and control of the level of Chinese herbal product specifications for the herb Amomum. Combined with IT technology, we analyze and modeling the experimental data to explore the generation of a practical, scientific and standardized method of Amomum organoleptic evaluation. The application of robust regression in the research to produce the prediction model achieved the classification forecast of Amomum product specifications.