Rough集在企业数据挖掘中的应用

企业信息系统中收集到的大量数据往往是不完备的,引入相似关系定义不完备信息系统的分明矩阵,采用分明矩阵的简化方法计算约简,对每个约简中的不完备数据,按照对象条件属性值相对决策属性取值的先验概率进行完备化,完备化后的信息系统生成了一组决策规则,这些规则不但存在冗余而且规则解释能力低,按照特征描述逐步挖掘规则并给出了挖掘所有最简决策规则的算法.整个过程中,通过客户信息资料描述一系列过程,验证了新的信息系统完备化方案以及挖掘算法的合理性和有效性.

Applications of Rough Set in Enterprise Data Mining

Amounts of data collected in enterprise information system are usually incomplete,the paper defines discernibility matrix by similarity r elation,uses method of simple discernibility matrix compute reduction,for inco mplete data of each reduction,completes via prior probability of object's condi tional attribute value towards decision attribute value,the completed informati on system generates a set of decision rules,these rules have not only redundanc ies but also low explainability,mine rules in the light of characteristic descr iption and gives algorithm for mining all the simplest decision rules.In the wh ole phases,the paper takes client datum information as an example to describe m ining process,validates that the new information system completion method and t he mining algorithm are rational and effective.