基于组合客观赋权法的进口茄衣质量评价

  目的  为建立更加合理的雪茄茄衣烟叶质量评价方法。  方法  以13个主要的进口茄衣样品为材料，测定了物理特性指标和化学成分指标，邀请专家对外观质量指标和感官质量指标进行打分，并对各指标数据进行了标准化处理。  结果  （1）4项一级指标的层次分析法主观赋权权重依次为外观质量（0.48），物理特性（0.22），化学成分（0.07），感官质量（0.23）；（2）组合了类间标准差法、CRITIC法和熵值法3种客观赋权后的二级指标权重的标准差较组合前提高；（3）结合了主客观赋权方法的综合赋权结果中，烟灰白度（0.0679）、残伤（0.0611）和支脉夹角（0.0444）指标权重系数较大，钾氯比值（0.0087）、干燥度（0.0095）和钾含量（0.0097）指标权重系数较小。  结论  该方法对一级质量指标采用主观法赋权，对二级质量指标采用3种客观法赋权，将主、客观权重的信息有效综合，较单一赋权更有效地反映出了各二级指标的重要程度。 

Quality evaluation of imported wrappers based on combined objective weighting method

  Objective  This study aims to establish a more reasonable quality evaluation method of cigar wrapper.  Method  By using 13 main imported wrapper samples as materials, their physical properties and chemical composition indicators were determined, and experts were invited to evaluate the appearance quality indicators and sensory quality. The indicators were scored and the data of each indicator was standardized.  Results  (1) The subjective weights of the four first-level indexes in the AHP included appearance quality (0.48), physical properties (0.22), chemical composition (0.07), and sensory quality (0.23); (2) The standard deviation of the weight sof the secondary index after the combination of inter-class standard deviation method, CRITIC method and entropy method was higher than before the combination; (3) In the comprehensive weighting results obtained by integrating subjective and objective weighting methods, the weight coefficients of indicators of smoke gray and white degree (0.0679), disability (0.0611) and branch angle (0.0444) were larger, and the weight coefficients of potassium-chloride ratio (0.0087), dryness (0.0095) and potassium content (0.0097) were smaller.  Conclusion  The subjective weighting method was adopted for the first-level quality indicators, and three objective weighting methods were adopted for the second-level quality indicators. By integrating both the subjective and objective weights, it is more reasonable to reflect the each second-level indicator weight than single weighting.