具有二次函数的多准则层次分类方法.pdfVIP

具有二次函数的多准则层次分类方法 王 坚 强 ( 中南大学商学院, 湖南长沙,410083,E-Mail:jqwang@) 摘要 提出了一种集成函数是二次函数且有训练集的多准则层次分类决策方法。该方法利用决策者对训练集的分类结 果（属于最高分类和不属于最高分类）构建非线性规划模型，然后采用一系列处理方法将非线性规划模型转换成线性规划 模型，求解线性规划得到训练集中各方案准则值的偏好值和相应参数，通过线性插值或样条插值得到每个方案在各准则下 的偏好值，并计算方案集中方案的与该分类一致性指标值和与低于该分类一致性指标值的差，比较得到方案是否属于该分 类，然后对不属于最高分类的训练集中的方案进行分类，并构建模型，继续上述过程，直到方案集中所有方案均分类为止。 最后实例说明该方法的有效性和可行性。 关键词 多准则决策 ；分类;偏好；二次集成函数；一致性指标 Hierarchical Discrimination Approach of Multi-criteria Classification With Quadratic Function WANG Jian-qiang （School of Business ， Central South University ，Changsha ， 410083 ，China） Abstract: A hierarchical discrimination method of multi-criteria classification with quadratic aggregation function is proposed. First, nonlinear programming is constructed by using classification (inside top category and outside top category) to reference set. Then nonlinear programming is converted into liner programming, and the preference values of alternative in the reference set and the relational coefficient is gained by solving the linear programming. And the preference values of alternative under the criteria and the category are obtained by using linear interpolation or spline interpolation, and the difference of concordance index of each alternative under category and lower the category are computed, and the alternatives being in the category or the lower category is acquired. Then the reference set, in which alternative belong to the lower top category, is classified, and programming model is constructed, and continue the above process until all alternative are classified. Finally, an example is given to explain the feasibility and availability of this method. Key words: multi-criteria decision-making; Classification; Quadratic aggregation function;