灰色聚类白化函数探析

Study on Grey Clustering Whitenization Weight Function

为进一步探析各类白化函数对综合评价结果的影响,先构建经典三角、修正三角1、修正三角2、经典聚类和指数型这五种不同类型的白化函数,再分别应用单一型数据、混合型数据、单指标单调型数据和跳跃型数据对这五种白化函数进行比较验证.结果显示两端等级的白化函数、零权重问题和等级区间宽度对综合评价结果影响较大.

To have a furtherstudy on the influence of the whitenization weight functions on the assessment of the synthetic grade, firstly five kinds of whitenization weight functionare established, including the classical triangular whitenization weight function, the improvement triangular whitenization weight function 1, the improvement triangular whitenization weight fuuction 2, the classical clustering whitenization weight function and the exponential whiteness function. Secondly, the superior and inferior of five kinds of whitenization weight function is validated by the usage of single type data, mixed type data, single index monotic type data and jumped type data. Lastly, itindicated that the whitenization weight function of the upperlower grade, zero-weight problem and the distance of the grade section have greatinfluence on the synthetic assessment.