Research Of Linguistic-valued Similarity-based Reasoning Method And Its Application

Knowledge representation,reasoning and classification are the first problem to be solved in intelligent application.There have existed a lot of problems with uncertain information in our daily life.However,in the representation and reasoning of uncertain knowledge and its negative,classical logic has not been satisfied with the development of knowledge processing.Most of the problems in the real world are imprecise,uncertain.Dealing with the problem of uncertain knowledge by using numerical calculation is limited.Therefore,based on the linguistic truth-valued lattice implication algebra,this paper makes a study on the value of the ten-element linguistic truth-valued lattice implication algebra and proposes linguistic-valued similarity-based reasoning method.The main research results are as follows:1.Firstly,according to the effect of the modal particles on the truth-valued,the modal particles can be divided into five categories by using the qualitative method.On this basis,the definition of the linguistic-valued "?" operation is obtained.Then achieving the linguistic-valued operation and study the properties of "?" operation.2.Analyzing properties of linguistic-valued and defining similarities and related properties between linguistic-valued.As the similarities between linguistic-valued,the method of utilizing modal language and distance calculated linguistic-valued similarity-based reasoning method is proposed.Moreover,the similarity of intuitionistic fuzzy linguistic-valued is proposed.It is a powerful extension as to linguistic-valued similarity-based reasoning.The linguistic-valued similarity-based reasoning applied to financial decision support system and fuzzy reasoning and logic-based agency,from knowledge representation,reasoning as computing speed and results will be compared,the results show that the reasoning knowledge representation linguistic-valued in terms of similarity closer to natural human language,easy to understand.linguistic similarity in inference computing speed computing equation simple reasoning square fewer steps so faster;results in terms of linguistic-valued similarity feel if reasoning is also side with the linguistic expression,fuzzy inference based on the accuracy of the agency's logic is poor in comparison.In practical applications,therefore,it should be based on different applications to choose different line of reasoning.It is recommended to use the linguistic-valued similarity-based reasoning to solve uncertain problems of missing information.3.Construct a knowledge-based value of ten-element linguistic-valued first-order predicate lattice implication algebra representation model combined with educational technology theory,the establishment of a network based on linguistic-valued similarity-based reasoning way to learn self-evaluation system.It is very important for network learners to develop their learning ability and improve the quality of learning by using this system.