Data mining tasks and methods: Classification: multicriteria classification

摘要: In this article we consider multicriteria classification, which differs from usual classification problems since it takes into account preference orders in the description of objects by condition and decision attributes. The well-known methods knowledge discovery do not use information about classification. It is worthwhile, however, to take as many practical involve evaluation on preference-ordered domains. To deal with propose a dominance-based rough set approach (DRSA). This different classical (CRSA) because domains attributes classes. Given partitioned predefined classes, new able approximate partition means dominance relations (instead indiscernibility used CRSA). approximation starting point for induction "if..., then..." rules. syntax these rules adapted represent orders. DRSA keeps best properties CRSA: only analyzes facts present data possible inconsistencies are corrected. Moreover, does need any prior discretization continuous-valued usefulness its advantages over CRSA presented real study risk business failure.