| Recently, it has become a hot topic to deal with large-scale, complex and synthetic data and extract interesting information from it.As is well known, it is effective to apply concept lattice theory in data analysis and knowledge acquisition. However, due to the completeness and precision of the concept lattice theory, it may have some shortcomings in deriving uncertainty knowledge and hence it cannot meet the requirement of more and more complex information acquisition. So, it is necessary for us to bring some new mathematical methods so as to extend the application of concept lattice theory in data mining. Rough set theory has just been demonstrated to be an effective mathematical tool in analyzing and handling incomplete and imprecise information. To the best of our knowledge, it has successfully been applied in many fields such as data mining. Motivated by these analyses, in this paper the rule acquisition theory of concept lattice is brought into the rough set theory, and it can effectively improve the ability of rule acquisition based on concept lattice.This paper mainly focuses on the interdisciplinary research of concept lattice and rough set theories in formal decision contexts, puts forward a new rule acquisition method, and makes a comparative study of rough set and concept lattice via rule acquisition. The main innovations are as follows:1) New rule acquisition methods are proposed in formal decision contexts, i.e., methods based on object-oriented concepts and attribute-oriented concepts. These new methods are compared with the original one based on Wille’s concept lattice. Moreover, theoretical analysis and experimental results show that the proposed methods are feasible and effective in decision-making problems.2) A comparative study between rough set and concept lattice is made via rule acquisition. Some useful relations and differences between these two theories are explored, and many meaningful results are obtained as well. |
PDF Full Text Request