针对不同主体评价不同客体存在不公平的缺陷,提出一种改进的评价算法.该算法首先根据同类主体不同成员给评价客体的评价分数的信息熵计算同类主体每一个成员的评价分数权重,并根据该权重改进TOPSIS方法,计算该类评价主体给出客体的评价分数距离该类主体给出的正、负理想解的相对贴近度.然后,根据每个客体在同类评价主体评价分数方差确定客体在该类评价主体的权重,并根据每一个客体在其被评价的不同类评价主体相对贴近度和权重计算每一名客体的总贴近度.根据所有客体的总贴近度进行排序.实例计算验证了本算法的有效性.%Aiming at the defect about the unfairness of the assessment of different object evaluated by different subject, an improved assessment algorithm was proposed. Firstly, the weight of subject for assessing the object for this class was calculated according to the information entropy of the scores about the each subject, and the TOPSIS method was improved with above weight of objects. Then, the relative closeness of positive and negative ideal solution of the all object for this class was calculated by the above improved TOPSIS method. Finally the weight of the object about the class was calculated according to the assessment scores given by the all subjects about the class. And the total relative closeness was calculated by the weight of the object about the class and the relative closeness about the same class. All subjects were sorted by the total relative closeness, and then the effectiveness of this proposed algorithm is validated.
展开▼