| Abstract: | Rough set theory provides a formal tool to deal with imprecise and incomplete information in terms of three valued logic. Reduct of attributes is the minimal subset of attributes without any loss of essential information. Using the reduct of attributes, we can reduce the computational burden on data mining algorithm and to incorporate only the important attributes in the rules, condense data and eliminate noise. Rough set approach to mining rules from databases was described, a new method for deriving optimal reduct and core of attributes was proposed. The heuristic algorithm ReductSearch for finding optimal reduct, and algorithm RuleBuild for mining rules from databases were given. |
