A novel approach to enhance the classification performances of grey relation analysis

Abstract

A Grey Relational Analysis (GRA) was a multiple criteria decision support approach in order to build a ranking and suggest a best choice on a set of alternatives. In this paper, a new Geometry Mean (GM) Pattern was proposed to enhance the applications of GRA.In this new GM pattern, the value could be calculated indirectly from the geometry mean of all attributes and the Attribute Impulse Factors and a different formula of GRG were defined. This study also classifies all of the existing GRG models into different patterns basically according to the kernel function Every pattern had the special characteristics and limitations, and they might be complemented to each other in the Grey Relational Analysis (GRA). In some particular applications were used to demonstrate that the New GM pattern derive a better satisfactory and effective evaluation than others. Finally, some useful conclusions deducted from discussions.

Keywords and phrases:

• Grey relational grade
• grey relational analysis
• pattern
• geometry mean pattern
• attribute impulse factor
• distinguishing coefficient