The Fuzzy Measure Based On The Knowledge And Its Application

Constructing a reasonable fuzzy measure is the key focus in many fields such asresource management and complex system optimization. The non-additivity of the fuzzymeasures can be intuitively explained as the correlation among elements, subadditivemeans that the union of two sup-parts plays a negative role, while superadditive meansthat the union of two sup-parts plays a positive role, which is coincided with thecharacteristics of intelligent decision, so fuzzy measures theories lay a theoretical basis fordifferent information fusion problem solving. The fuzzy measures theories are relativelymature, but the methods of determining fuzzy measures still lack systematic operationmechanism, and it is the bottleneck restricting the application. In this paper, ourcontributions are as follows:Firstly, aiming at interactions between attributes (indexes), after analyzing featuresand deficiencies of existing fuzzy measures, we give the concept of core samples set underthe background of decision system, following which we discuss the dependence of thecore samples set and conditional attribute. Furthermore, combining with structure featuresof fuzzy measures, we give a method based on core samples set to measure the correlationbetween attributes. Finally, we analyze its characteristics and performance through aconcrete example. The results show that the measure can describe effectively correlationbetween attributes, and possess good interpretability. So, it can be applied in many fieldssuch as synthetic decision-making and information management.Secondly, since lots of decision-making problems is the summary of existingknowledge, we can measure the importance of condition attributes sets by using the changeof the core samples set as a strategy of. After analyzing features and shortage of existingfuzzy synthetic evaluation methods, we propose a fuzzy synthetic evaluation model basedon the knowledge system by a metric method of attribute correlation under the backgroundof decision system. Then we analyze its characteristics and performance through thesynthetic evaluation of university teachers’ teaching quality. Finally, we compare theresults of the fuzzy synthetic evaluation based on the Choquet fuzzy integral with theresults of the fuzzy synthetic evaluation based on the Sugeno fuzzy integral.