The Design And Implementation Of The Belief Nonrevision Prototype System Interpreter

The knowledge in daily life,which is often uncertain,inconsistent and incomplete,is referred to as belief.How to deal with inconsistencies among beliefs is a key issue in stimulating reasoning process of computers.Currently,to update beliefs revision,such problems like:losing the important information,resulting in a conclusion which is useless,conclusions which are difficult to choose,exist inevitably.The non-revision method based on non-revision,however,has resolved the above problems.Research content of this paper is to make improvements on the basis of non-revision reasoning method with belief in the first order predicate logic,using unit order method to reason.According to user's Chinese input,transform the clauses of the knowledge base and use the unit order method to reason;when detecting the inconsistent knowledge,select in accordance with the confidence level.Finally,achieve a conclusion which does not show to the user the reasoning process,so the user is suspicious of the authenticity of the conclusions.In order to give the user the final interpretation of the conclusion,the Belief Non-revision Prototype System interpreter is designed.In the process of attribution,the set of clauses into beliefs is constructed,adopting the principle of reasoning to track the process,the interpreter reverses the reasoning process by traversing the belief clauses.The knowledge used to get the conclusion is connected,the reasoning process of the conclusion is constructed,and the user can analyze the accuracy and validity of the reasoning conclusion based on the reasoning process.