混合属性广义灰靶决策方法的模式识别功能研究

On the Function of Pattern Recognition of Mixed Attribute-Based Generalised Grey Target Decision Method

为研究混合属性广义灰靶决策方法的模式识别功能,根据其具有靶心指标可变动,决策依据(靶心距)具有开放性、可扩充的特点提出了可用于模式识别的理论框架。该理论框架首先确定标准模式的特征及待识别目标,分别作为靶心指标向量和决策方案;其次,化各类混合属性值为二元联系数,并视为指标向量;然后,基于指标向量计算各待识别目标向量与靶心向量的接近度;最后,根据各待识别目标的综合接近度确定待识别目标的相应模式。应用分析表明,投资企业y_1和y_2与模式S_3和S_2的最小综合接近度分别为0.427 0和0.237 5;因此,企业y_1与y_2分别选择投资模式S_3和S_2是适宜的。

The function of pattern recognition of the generalised grey target decision method for mixed attributes is investigated.Thus,the theory framework of pattern recognition is proposed for the method which has its characteristics:variable target centre index,open and extensible decision-making basis(target centre distance).The theory framework goes as follows.First,the characteristics of standard patterns(regarded as the target centre index vectors)and the researched objects are determined;second,all mixed attribute values are converted into binary connection numbers(viewed as the index vectors);third,the generalised target centre distances between the researched objects and the target centre indices are calculated;and finally,the researched objects are recognised as some special patterns relying on the comprehensive weighted proximities.The application shows that the two investment enterprises y1 and y2 could select their suitable investment patterns S3 and S2 respectively(corresponding to their comprehensive proximities as 0.4270 and 0.2375 respectively).