提出了一种基于加权模糊规则的新的推理机制,并将这组加权模糊规则及相应推理机制映射成了一个模糊神经网络,其中加权模糊规则中的(局部和整体)权重恰好对应于神经网络的连接权.通过新改进的BP算法训练此神经网络后,可学习得到加权模糊规则的权重近似优值.模拟实验说明,训练后得到权重,用本文提出的新的推理机制可以一定程度上提高推理精度.
[1] YEUNG D S, TSANG E C C. Weighted fuzzy production rules [J]. Fuzzy Sets and Systems 1997, 88.
[2] YEUNG D S, TSANG E C C. A multilevel weighted fuzzy reasoning algorithm for expert systems [J]. IEEE Transactions on Systems Man and Cybernetics 1998, 28(02).
[3] TSANG E C C, YEUNG D S, WANG X A. Learning weights of fuzzy production rules by a max-min nural netwotk [A]. Romana, Bucharest 2001.
[4] Tsang E.C.C., Yeung D.S., Lee J.W.T., Huang D.M., Wang X.Z.. Refinement of generated fuzzy production rules by using a fuzzy neural network [J]. IEEE transactions on systems, man, and cybernetics, Part B. Cybernetics: A publication of the IEEE Systems, Man, and Cybernetics Society 2004, 1(1).
[5] YUAN Y, SHAW M J. Induction of fuzzy decision trees [J]. Fuzzy Sets and Systems 1995, 69(02).