参数化模糊联想记忆网络的学习算法

Learning algorithm for parameterized fuzzy associative memory

基于Dubois提出的带参数ξ的t-模Tξ,提出了一种参数化的广义模糊联想记忆网络Max-TξFAM。由于Tξ中参数ξ的作用,在应用中Max-TξFAM有更强的可调性和灵活性。接着利用Tξ的伴随蕴涵算子,提出了Max-TξFAM的一种有效学习算法。从理论上严格证明了,只要Max-TξFAM能完整可靠地存储所给的多个模式对,则所提出的学习算法一定能找到使得网络能完整可靠存储这些模式对的所有连接权矩阵的最大者。最后,用实验说明了所提出的学习算法的有效性。

Based on fuzzy composition of maximum operation and a t-norm Tξ with a parameter ξ proposed by Dubois,a parameterized general fuzzy associate memory Max-Tξ FAM is presented in this paper.By adjusting parameter ξ,the Max-Tξ FAM has good adaptability and flexibility in practice.Taking advantage of the concomitant implication operator of Tξ,a simple effective leaming algorithm is proposed for the Max-Tξ FAM.h is proved theoretically that,for arbitrary given training pattern pairs,if the Max-Tξ FAM has ability to store them reliably and completely,then the proposed learning algorithm can find the maximum of all connected weight matrices which can ensure that the Max-Tξ FAM stores reliably these pattern pairs.Finally an experiment is given to illustrate the effectivity of the presented learning algorithm.