Attribute Reduction Of Concept Lattice Based On Formal Context Transformation And Consistent Decision Concept

The concept lattice acts as an effective tool for cognitive learning,and its attribute reduction benefits the optimization processing of concept extension and intension.This thesis discusses the consistent attribute reduction based on formal context transformation and the concept consistent attribute reduction based on consistent decision concept.The specific research contents are as follows.(1)By referring to the classical attribute reduction of concept lattice in formal context,new attribute reduction of concept lattice in decision formal context is proposed by transformation.At first,consistent sets are defined in decision formal context,and consistent reduction and their algorithm are proposed for concept lattice;moreover,connections between consistent attribute reduction and three-way attributes are acquired.Then,relationships of consistent attribute reduction regarding existing strong-consistent and weak-consistent attribute reduction are investigated;thus,consistent attribute reduction exhibit the strict strength in terms of attribute reduction of concept lattice,and this conclusion motivates heuristic reduction algorithms from the strong reduction to the weak.(2)The definition of consistent decision concept is given,and the concept consistent attribute reduction based on consistent decision concept is proposed.And study the relationship between the new concept strongly consistent attribute reduction and the existing strongconsistent attribute reduction,and obtains the two-way relationship between the concept strongly consistent attribute reduction and the strong-consistent attribute reduction.A concept strongly consistent attribute reduction can be constructed from the inside of a strong-consistent attribute reduction,and a strong-consistent attribute reduction can be derived from the union of all the concept strong-consistent attribute reductions,and the corresponding algorithm is designed.Similarly,the two-way relationship between the concept weakly consistent attribute reduction and weak-consistent attribute reduction is given.In a word,this thesis mainly proposes two attribute reduction methods.The relationships between the two kinds of reductions and strong-consistent and weak-consistent reductions is studied and verified respectively.The relevant theorems and algorithms are verified by examples.The two attribute reduction methods proposed can be applied in the context of multiple decision forms and can meet the requirements of practical applications.