Minimum cost attribute reduction in incomplete systems under decision-theoretic rough set model

Decision-theoretic rough set model has aroused extensive attention of recent years during the development of probabilistic rough set. Many studies have defined different reducts in different rough set model. However, few studies report the attribute reduct in incomplete information systems under the decision-theoretic rough set model. By considering three classical extended rough set models for incomplete systems, this paper present the new definition for the attribute reduct in decision-theoretic rough set model. The objective of the reduct is to seek for the set of attributes which have the minimal subset, simultaneously the subset has the minimum decision cost. A heuristic reduction approach is also designed and the results of the experiment show the high efficiency of our approach.