灰色关联分析中分辨系数取值的定量研究

Quantitative Study on Value of Distinguishing Coefficient in Grey Correlation Analysis

灰色关联分析模型作为一种有效的定性比较方法应用于经济、生态、决策方法等多个领域。在灰色关联度计算中,分辨系数ρ的不同取值将影响关联度排序的结果。通常研究中依靠经验对ρ取值,使ρ=0.5;或者根据序列的波动情况定性地对ρ取大值或小值。为增大关联度之间的分辨率,文章采用了数学推导的方式对ρ进行定量取值。实例证明,这种取值方式能提升不同因素关联度之间的分辨率,使比较更准确。

As an effectively qualitative and comparative method, grey correlation analysis model has been applied in many fields such as economy, ecology, decision-making and so on. In the calculation of grey relation grade, different values of distinguishing coefficientρwill affect the results of correlation degree ranking. Usually in studies, the value ofρis determined by experience, making ρ equal to 0.5;or the big value or small value of ρ is qualitatively taken according to the fluctuation of the sequence. In order to increase the resolution between correlation degrees, the paper uses the method of mathematical deduction to quantify the value of ρ. The example proves that this method can improve the resolution of correlation degree of different factors and make the comparisons more accurate.