Theory and Practice on Information Granule Matrix

In this paper, a new framework called information granule matrix is suggested to illustrate a given granule sample for showing its information structure. The new framework does not any extra condition but the observations. An information granule matrix can be turned into a fuzzy relation matrix for fuzzy inference. The concept of information granule matrix is firstly formulated. Then information granule matrix is shown by a simple example and discussed from the meaning of mechanism. To display the advantage of the new framework, it is compared with some existed methods. We also use our suggested framework to illustrate the relationship between earthquake magnitude M and isoseismal area S. The result shows that the new model is better than both Linear Regression and BP Network.

[1]  Yong Shi,et al.  Towards Efficient Fuzzy Information Processing - Using the Principle of Information Diffusion , 2002, Studies in Fuzziness and Soft Computing.

[2]  Claudio Moraga,et al.  A diffusion-neural-network for learning from small samples , 2004, Int. J. Approx. Reason..

[3]  Peter Raeth,et al.  Book review: Fuzzy Engineering by Bart Kosko (Prentice Hall, 1997) , 1998, SGAR.

[4]  Chongfu Huang,et al.  Some concepts and methods of information granule diffusion , 2005, 2005 IEEE International Conference on Granular Computing.

[5]  Alberto Tesi,et al.  On the Problem of Local Minima in Backpropagation , 1992, IEEE Trans. Pattern Anal. Mach. Intell..

[6]  Bart Kosko,et al.  Fuzzy Engineering , 1996 .