Rule approximation in metric spaces

The classical fuzzy rule interpolators work in Euclidean spaces where the new fuzzy value can be generated from the training values. The paper investigates the case when the fuzzy values are defined over the general metric spaces. In this case a classification process is used to approximate the requested value. The paper introduces a base method for this classification process.

[1]  László T. Kóczy,et al.  Approximate reasoning by linear rule interpolation and general approximation , 1993, Int. J. Approx. Reason..

[2]  S. Kovács.  Interpolation-based Fuzzy Reasoning as an Application Oriented Approach , 2005 .

[3]  László Kovács,et al.  Multi-Layered Semantic Data Models , 2009, Encyclopedia of Artificial Intelligence.

[4]  László T. Kóczy,et al.  A generalized concept for fuzzy rule interpolation , 2004, IEEE Transactions on Fuzzy Systems.

[5]  Zsolt Csaba Johanyák,et al.  A BRIEF SURVEY ON FUZZY SET INTERPOLATION METHODS , 2006 .

[6]  C. Marsala,et al.  Interpolative reasoning based on graduality , 2000, Ninth IEEE International Conference on Fuzzy Systems. FUZZ- IEEE 2000 (Cat. No.00CH37063).

[7]  László T. Kóczy,et al.  Stability of interpolative fuzzy KH controllers , 2002, Fuzzy Sets Syst..