Research On Formal Concept Based On Attribute Classification

In 1982,Professor Wille first proposed the theory of formal concept analysis,which is a tool for data analysis and rule extraction from formal context.For the theory of formal concept analysis,the existing researches mainly focus on the acquisition of formal context knowledge and the calculation of formal concept.The formal context is the data source of formal concept analysis,and the calculation of formal concept is the data structure of formal concept analysis.This paper focuses on two aspects:the knowledge of formal context and.the calculation of formal concept.Based on the idea of granular computing,this paper studies the attribute reduction,attribute classification and dynamic calculation of formal concept,and the results obtained are shown as follows.(1)This paper proposes a formal concept attribute reduction algorithm based on attribute classification relation.Firstly,there are some obvious shortcomings in the existing formal concept attribute reduction algorithms,such as the time complexity of computing attribute reduction is high;the attribute equivalence class and attribute reduction are calculated separately,there is redundant computation;the conversion of formal context knowledge to the covering knowledge increases the cost of system storage and so on.In order to solve these problems,two heuristic operators are defined in this paper,and the relationship between attributes is calculated.Then,this paper proposes an attribute reduction algorithm based on attribute classification relation.This algorithm can effectively reduce the computational time complexity,and reduce the redundant computation and the system storage cost,and improves the computational efficiency of the attribute reduction.Finally,the validity of the formal concept attribute reduction algorithm based on attribute classification relation is verified by examples and experiments.(2)In this paper,we propose a hierarchical formal concept analysis model based on attribute classification and a dynamic construction algorithm based on attribute classification.First of all,in the traditional formal concept analysis,the attributes are single granularity and single level.However,the attribute classification of tree structure is a common problem in reality.Then,this paper proposes a multi-level formal concept model based on attribute classification,and analyzes the related properties of formal concept in different levels of generalization space,and puts forward a method of attribute generalization and refinement based on formal concept analysis,in which attribute generalization simplifies concept and attribute refinement improves accuracy.On the basis of this,a dynamic formal concept construction algorithm is proposed based on the change of attribute classification level,the algorithm adopts the self-learning method to use the existing knowledge,which not only inherits the advantages of the existing incremental algorithm,but also can deal with the problem of its own data dynamic changes of the formal concept.Finally,the effectiveness of the proposed method is illustrated by examples and simulation experiments.