Research And Application Of Fuzzy Expert System Based On Vague Sets

Expert system is a most active embranchment of artificial intelligence and knowledge engineering.Because of being inaccuracy,there are the fuzzy expert systems that handle inaccuracy.The traditional fuzzy expert systems handle inaccuracy of knowledge with fuzzy sets.Gau and Buehrer give vague sets that describe problem of fuzzy information and data.Approximate reasoning is the kernel of the fuzzy expert system. At present, Compositional Rule of Inference is often used in approximate reasoning mechanism of the fuzzy expert system. A ameliorative measures of similarity is put forward after we discuss merit and shortcoming of the traditional measure of similarity.A ameliorative method of approximate reasoning is proposed based on the ameliorative measure of similarity.We analyse application of the ameliorative method of approximate reasoning on conditions that there is one rule or many rules. In addition, the value of vague set that we get based on all methods of similarity is less than 1 or is equal to 1,so the vague set of the result can not reflect the trend of increase or reducement of vague set.A method of approximate reasoning based on capability to vague set is put forward,the value of vague set is got by capability of vague set.We analyse application of this ameliorative method of approximate reasoning on conditions that there is one rule or many rules also.At last,a fuzzy consultation system on student choosing department based on the ameliorative measures of similarity of approximate reasoning is developed.This system may get degree of matching of student choosing certain department of university based on 3 factors including evaluation of examination, interest and adaptability of later job.Vague sets describe fuzzy information of three factors better,so the consultation result accord with reality.