The Research On Incremental Learning Method In Cognitive Concepts Based On Attribute Topology

Formal concept is one of the core components of Formal Concept Analysis(FCA),and currently has become an important research aspect of Artificial Intelligence related disciplines.The emergence of formal conceptual computing methods promotes the development of FCA in various fields such as knowledge discovery,data mining,and cognitive computing.As a new type of formal representation of the context,the attribute topology(AT),which combines the advantages of information science,describes the correlation information between attributes directly,which attracts more and more attention.This paper is based on the basic theory of FCA and AT.The following studies were carried out:Firstly,based on Depth First Search and Object Topology(OT),an incremental conceptual computing algorithm is proposed.Through the discussion of the influence of the new object to the original concept structure,the classification of the new object is realized.On this basis,a collapse algorithm of OT is proposed,which makes the learning process more focused,and reducing the complexity of the learning process.Finally,a topology updating algorithm based on the collapse algorithm is combined with the Depth First Search algorithm to provide an incremental conceptual computing algorithm,Recursive Depth First Formal Concept Search(RDFFCS),as a basis for the subsequent conceptual computation and cognitive learning.Secondly,we propose a conceptual computing algorithm that originated from AT but not limited to AT,an Incremental Concept Computing Algorithm Based on Concept Tree.Firstly,the incremental updating algorithm based on the Recursive Depth First Search is summarized and the influence of the path update on the concept generations discussed.By analyzing the principle of path updating in the search process,the rules between the path and the path tree are strengthened,and the association between the concept tree and the concept update is studied.Finally,the Incremental Concept Computing Algorithm Based on Concept Tree is obtained.Finally,the incremental concept computing algorithm proposed in this paper is applied to the conceptual cognitive learning system of professional courses,and experimental test of eight semesters in four academic years respectively.Through the analysis,it is concluded the feasibility and correctness in the cognitive learning system,and discussed the significance of the method in practical application.