Modification of a fuzzy partition often leads to the change of fuzziness of a fuzzy system. Researches on the change of fuzzy entropy of a fuzzy set, responding to shape alteration of membership function, therefore...Modification of a fuzzy partition often leads to the change of fuzziness of a fuzzy system. Researches on the change of fuzzy entropy of a fuzzy set, responding to shape alteration of membership function, therefore, play a significant role in analysis of the change of fuzziness of a fuzzy system because a fuzzy partition consists of a set of fuzzy sets which satisfy some special constraints. This paper has shown several results about entropy changes of a fuzzy set. First, the entropies of two same type of fuzzy sets have a constant proportional relationship which depends on the ratio of the sizes of their support intervals. Second, as for Triangular Fuzzy Numbers （TFNs）, the entropies of any two TFNs which can not be always the same type, also, have a constant proportional relationship which depends on the ratio of the sizes of their support intervals. Hence, any two TFNs with the same sizes of support intervals have the same entropies. Third, concerning two Triangular Fuzzy Sets （TFSs） with same sizes of support intervals and different heights, the relationship of their entropies lies on their height. Finally, we point it out a mistake that Chen＇s assertion that the entropy of resultant fuzzy set of elevation operation is directly to that of the original one while elevation factor just acts as a propartional factor. These results should contribute to the analysis and design of a fuzzy system.展开更多
Fuzzy entropy has been widely used to analyze and design fuzzy systems, and many fuzzy entropy formulae have been proposed. For further in-deepth analysis of fuzzy entropy, the axioms and some important formulae o...Fuzzy entropy has been widely used to analyze and design fuzzy systems, and many fuzzy entropy formulae have been proposed. For further in-deepth analysis of fuzzy entropy, the axioms and some important formulae of fuzzy entropy are introduced. Some equivalence results among these fuzzy entropy formulae are proved, and it is shown that fuzzy entropy is a special distance measurement.展开更多
This paper presents a method of modeling a fuzzy system with fuzzy and nonlinear border,obtaining systematic structure by clustering analysis,applying BP network (BPN) to generate rule bases antecedent function and us...This paper presents a method of modeling a fuzzy system with fuzzy and nonlinear border,obtaining systematic structure by clustering analysis,applying BP network (BPN) to generate rule bases antecedent function and using RBF network (RBFN) to approximate each rules conclusion function not only because of efficient capability of approximation nonlinear function of BPN and RBFN but also because of quickness of training speed of RBFN. In addition,structure design and training of relevant networks are discussed in detail. Finally,the structure optimization and overstudy of RBFN are discussed.展开更多
Let X = （x1 ,x2 ,…… ,xn ） and F（X） be a fuzzy set on a universal set X. A new def＇mition of fuzzy entropy about a fuzzy set A on F（X）, e^＊＇ , is defined based on the order relation ＂≤＂ on [ 0,1/2 ]^n. It...Let X = （x1 ,x2 ,…… ,xn ） and F（X） be a fuzzy set on a universal set X. A new def＇mition of fuzzy entropy about a fuzzy set A on F（X）, e^＊＇ , is defined based on the order relation ＂≤＂ on [ 0,1/2 ]^n. It is proved that e^＊ is a σ-entropy under an additional requirement. Besides, some entropy formulas are presented and related properties are discussed.展开更多
基金The National Natural Science Foundation of China(No.60474022)
文摘Modification of a fuzzy partition often leads to the change of fuzziness of a fuzzy system. Researches on the change of fuzzy entropy of a fuzzy set, responding to shape alteration of membership function, therefore, play a significant role in analysis of the change of fuzziness of a fuzzy system because a fuzzy partition consists of a set of fuzzy sets which satisfy some special constraints. This paper has shown several results about entropy changes of a fuzzy set. First, the entropies of two same type of fuzzy sets have a constant proportional relationship which depends on the ratio of the sizes of their support intervals. Second, as for Triangular Fuzzy Numbers （TFNs）, the entropies of any two TFNs which can not be always the same type, also, have a constant proportional relationship which depends on the ratio of the sizes of their support intervals. Hence, any two TFNs with the same sizes of support intervals have the same entropies. Third, concerning two Triangular Fuzzy Sets （TFSs） with same sizes of support intervals and different heights, the relationship of their entropies lies on their height. Finally, we point it out a mistake that Chen＇s assertion that the entropy of resultant fuzzy set of elevation operation is directly to that of the original one while elevation factor just acts as a propartional factor. These results should contribute to the analysis and design of a fuzzy system.
文摘Fuzzy entropy has been widely used to analyze and design fuzzy systems, and many fuzzy entropy formulae have been proposed. For further in-deepth analysis of fuzzy entropy, the axioms and some important formulae of fuzzy entropy are introduced. Some equivalence results among these fuzzy entropy formulae are proved, and it is shown that fuzzy entropy is a special distance measurement.
文摘This paper presents a method of modeling a fuzzy system with fuzzy and nonlinear border,obtaining systematic structure by clustering analysis,applying BP network (BPN) to generate rule bases antecedent function and using RBF network (RBFN) to approximate each rules conclusion function not only because of efficient capability of approximation nonlinear function of BPN and RBFN but also because of quickness of training speed of RBFN. In addition,structure design and training of relevant networks are discussed in detail. Finally,the structure optimization and overstudy of RBFN are discussed.
基金The National Natural Science Foundation of China (No.60474022) and the Fundamental Science Foundation of Southwest Jiaotong University (No.2004B08)
文摘Let X = （x1 ,x2 ,…… ,xn ） and F（X） be a fuzzy set on a universal set X. A new def＇mition of fuzzy entropy about a fuzzy set A on F（X）, e^＊＇ , is defined based on the order relation ＂≤＂ on [ 0,1/2 ]^n. It is proved that e^＊ is a σ-entropy under an additional requirement. Besides, some entropy formulas are presented and related properties are discussed.
