多标准决策表中发现概率规则的变精度粗糙集方法

用优势关系代替不可分辨关系,本文提出了一种可以从多标准决策表中获取概率规则的扩展变精度粗糙集模型,该模型能够处理多标准决策表中可能的不一相容性,获取由偏好对象组成的概率决策规则集,并导出事例决策系统的偏好模型。研究结果表明:基于优势关系,从多标准决策表中获取的最小概率规则集,使用的条件数量较少,且导出规则的数量较少、较强。

Variable Precision Rough Sets Methodology for Probabilistic Rules Discovery from a Multi-Criteria Decision Table

An extension of variable precision rough sets model is proposed by dominance relation instead of indiscernibility relation,which can derive probabilistic rules from multi-criteria decision tables.The model can deal with possible inconsistencies in multi-criteria decision tables,obtain probabilistic decision rules made up of preference actions.Based on these rules,a preference model from example decision systems is inferred.Research results show that the minimal set of probabilistic decision rules obtained from a multi-criteria decision table by dominance relation has a smaller number of stronger rules and a smaller number of conditions.