为解决指标体系构建中的指标约简与权重分配问题,利用灰色理论分析条件指标与决策指标之间以及不同条件指标之间的关联性,在此基础上给出指标重要度和指标之间影响度的量化定义,借助上述定义,提出条件指标去重叠化方法,设计基于灰关联分析的指标约简与权重分配算法。与其它指标约简算法相比,该算法能够同时处理离散型决策表和连续型决策表以及二者同时存在的混合决策表,能够基于条件指标绝对重要度进行权重分配。
To reduce indexes and distribute weight when building indexes system,the correlation among condition indexes and the correlation between condition indexes and decision indexes were analyzed using gray theory.On this basis,the quantitative definitions of index importance degree and index effect degree were given.With these definitions,a method for eliminating overlap among condition indexes was proposed and an algorithm of indexes reduction and weight distribution was designed.The algorithm can be used for continuous decision table,discrete decision table and hybrid decision table,and the weight was distributed based on the condition index absoluteness importance degree.