基于粗糙集理论不完备信息系统的数据挖掘

A Data Mining Method Based on Rough Sets Theory for Incomplete Information System

提出了一种基于推广的粗糙集理论直接在不完备信息系统上进行数据挖掘的方法 ,并给出了该方法的算法和实例 .该方法利用粗糙集理论直接对不完备信息系统进行知识约简 ,然后根据获得的约简集建立知识层次树 ,利用规则的支持度阈值s0 和置信度阈值c0 从知识层次树的压缩搜索空间中提取不完备系统的规则集 .该方法保持了原始数据和数据挖掘所获得的知识的真实性 .另外 ,还提出了知识规则的上、下支持度 ,上、下置信度 ,规则粗糙度等概念 ,以便指导用户更好地利用数据挖掘所获得的知识 .

A data mining method based on rough sets theory for incomplete information system is proposed. An algorithm and an instance for this method are given here,also. This method employs rough sets theory by which incomplete information system is reduced. Layered knowledge trees are constructed with sets that are gained by reducing incomplete information system. With thresholds of confidence,c_0,and support,s_0,rule sets of incomplete system are abstracted from the compacted special of the trees. This method keeps the truth of the raw data and the knowledge gained by mining data. Moreover,concepts,such as upper support,lower support,upper confidence,lower confidence,roughness of rules,which can guide users to make good use of knowledge mined from data,are proposed,also.