在知识发现、数据挖掘等研究领域中,粒度计算有着潜在的应用。针对决策表,探讨了其内在的粒度思想,研究了属性个数增减、属性取值改变对知识粒度变化的影响,并建立了知识粒度与规则确信度之间的动态关系,为决策表进行决策分析提供了科学依据。这些结果将有助于粒度计算和决策理论的研究。
Granular computing is potentially used in knowledge discovery and data mining etc. Based on decision table, the paper discusses its internal granular essence, researches the relationship between knowledge granulation and the number of attributes or the change of attribute value, and establishes the dynamic relation between knowledge granulation and the credibility of rules. These results will be helpful for studying for granular computing and decision system theory.