Study on the Evaluation Method of Flue-cured Tobacco Sensory Characters
|School||Henan Agricultural University|
|Keywords||Flue-cured tobacco Sensory quality Fuzzy comprehensive evaluation Extension analysis Attribute recognition theoretical model|
Although sensory evaluation of flue-cured tobacco had certain subjectivity, it was still the main basis of tobacco quality identification, leaf group blending, screening the proportion of essence and liquid formula. It could directly make an accurate judgment of tobacco quality by means of examining sensory indexes, then sensory quality evaluation emerged as the times require.Most common evaluation method of sensory quality was index sum method. It multiplied different evaluation index score with corresponding weights, incoming plus and namely for final evaluation value. This method was simple and easy to realize. But the sensory quality of flue-cured tobacco depended not only on weighted sums of indexes, but was more influenced by few limiting factors. If only using index sum method to calculate the sensory quality, it would hide the influence of limiting factor on the quality of flue-cured senses, so the accuracy of the evaluation results would decrease obviously.As the importance of different indexes was not the same, different indexes should be determining weights when using comprehensive evaluation. Method of determining weights in sensory quality evaluation was relatively single. The subjective weighting method of analytic hierarchy process and expert consultation were mostly used. Although this method embody the wish of the experts relatively superior, it ignored the character of data.According to the problems mentioned above, the fuzzy comprehensive evaluation, matter-element extension analysis and attribute recognition theoretical model were used to evaluate sensory quality of flue-cured tobacco. Combination weighting method was adopted to weight idexes. Sensory quality datas of middle leaves from 28 counties and districts in Sichuan province were used as an example to compare and study.In the evaluation process, five main indexes of sensory quality were determined based on analysis of the previous researches, including: aroma quality, amount of aroma, offensive odor, aftertaste and irritation. The sensory quality evaluation standard was also established. The subjective and objective weight of indexes were determined respectively by means of scale expansion-AHP and entropy weight. Combination weight was determined through square normalization. The weights of aroma quality, amount of aroma, offensive odor, aftertaste and irritation were 0.2496,0.3308,0.2229,0.1115 and 0.0852 respectively.According to the relationship of total point and sensory quality indexes, half liter gradient function was determined to be the membership function in fuzzy comprehensive evaluation. The fuzzy mixture was based on weight average operator. The evaluation results showed that: within 28 counties and districts, 3 were ranked in“better”rank and the“good”rank included 11 countries, which mainly distributed in Liangshan. The results of attribute recognition theoretical model was: within 28 counties and districts, 1 was ranked in“better”rank and the“good”rank included 13 countries.Based on relational function, Single indexes relation and comprehensive relation were calculated, and then the matter-element extension evaluates assessment indexes, the result showed that 1 was ranked in“better”rank and the“good”rank included 5 countries.The results of using three different evaluation methods were almost same, specifically as follows:Huili,Jingyan,Puge,Butuo,Yuexi of Liangshan,Yanbian of PanZhiHua,Xuyong of Luzhou,Pingshan,Wangcang ,Daxian, were completely same.