Decision Rule Mining in Rough Set Theory

Rough se theory(RST) has two formats, abstract and table formats. In this article, abstract format is hardly touched. The table format, by definition, is a theory of extensional relational databases. However, their fundamental goals are very different. RST has focused on data analysis and mining, while database has focused on data processing. In RST, a relation, which is also known as information table, is called a decision table (DT), if the attributes are divided into two disjoint families, called conditional and decision attributes. A tuple in such a DT, is interpreted as a decision rule, namely, the conditional attributes functionally determine decision attributes. A sub-relation is called a Value Reduct, if it consists of a minimal subset of minimal length decision rules that has the same decision power as the original decision table. RST has the following distinguished theorem.