中文摘要：

为了降低利用D-S证据组合结果进行决策的风险，提出了一种基于集合属性和优先度的证据决策方法，将决策问题分解成精细信度区间的构造和比较两个层面。在构造层面引入集合的不确定性测度和焦元间属性支持度，获取基元命题的精细信度区间值；在比较层面引入优先度评价各命题的精细信度区间，然后在优先度排序的基础上构建了证据决策模型。该方法的特点在于以区间值作为决策依据，充分利用了信度区间所蕴含的信息。实验分析结果表明，该方法是合理、有效的，克服了其它单点值D—S证据决策方法所存在的误决策或不做决策的问题。

英文摘要：

To reduce the risk of decision-making using D-S evidence fusion results, a novel decision rule based on set attribute and preference degree is proposed. The method divides the decision problem into two levels, construction level and evaluation level of belief interval. At the construction level, by means of the uncertainty measure of set and support degrees between two focal elements, the refined belief interval of base elements is obtained. At the evaluation level, the preference degree is used to estimate and rank the defined belief interval. Then the decision-making model is constructed on the basis of the preference degree order. The method is characterized by making full use of interval information as a basis for decision-making. Comparing with other endpoint value decision rules, the numerical examples show that proposed method can avoid the wrong decision-making effectively.