An Attribute Reduction Approach Based On Three-way Fuzzy Linguistic Concept Lattice

With the rapid development of information society,a lot of vague or uncertain information appears in daily life.Uncertainty information processing is an important research content in the field of artificial intelligence.Formal concept analysis is an efficient data processing tool,which has been widely used in many fields such as pattern recognition,information retrieval,and artificial intelligence.However,it has certain limitations in the processing of fuzzy information and linguistic values.Therefore,this paper combines the three-way concept lattice with linguistic values,and proposes three-way fuzzy linguistic concept lattices and attribute reduction method.The main research results are shown as follows:1.In order to simplify the fuzzy linguistic formal context and improve the efficiency of calculation,we propose the coarse-grained objects similarity by granular computing with the linguistic value.In order to coarsen the fuzzy linguistic formal context,a similarity threshold is proposed to realize the simplification of the formal context.That is,the coarse-grained objects fuzzy linguistic formal context.Users can change the similarity threshold according to their own needs to obtain formal contexts in different granularities.The coarse-grained objects fuzzy linguistic formal context is applied to the teaching evaluation example so as to solve the problem of formal context which is too complicated.Subsequently,based on the related knowledge of the adjacency matrix and the incidence matrix,an attribute topological graph is generated to describe the internal relationship between attributes and objects in the concept lattice.2.The three-way concept analysis can divide information into acceptance,rejection and non-commitment,and has a wide range of applications in real life.In order to express linguistic value information in various aspects,we combine three-way concept analysis with fuzzy linguistic information to construct three-way fuzzy linguistic concept lattices.The linguistic threshold is proposed and the positive and negative operators are derived,so as to construct the object-induced three-way fuzzy linguistic operators and the attribute-induced three-way fuzzy linguistic operators.Then we construct the object(attribute)-induced three-way fuzzy linguistic concept lattices to deal with fuzzy information comprehensively.At the same time,in order to express the construction process more clearly,we propose the construction method of three-way fuzzy linguistic concept lattices.3.On the basis of three-way fuzzy linguistic concept lattices,the attribute reduction of three-way fuzzy linguistic concept lattices is deeply researched and discussed.Based on the discinibility matrix and discinibility function,we propose the object(attribute)three-way fuzzy linguistic concept lattices discinibility matrix and discinibility function in order to reduce and divide attributes.In order to specifically describe the reduction process,an attribute reduction algorithm is proposed and data comparative analysis is carried out to verify its accuracy and advantages.The algorithm can be used in artificial intelligence,machine learning and other fields.