论文部分内容阅读
提出一种结构等价型模糊神经网络的学习算法．等价型神经网络根据模糊系统的推理规则，决定等价的神经网络结构参数，因而网络结构特殊．采用的学习算法是用误差反传对局部节点的权值进行调整，收敛速度快．实验表明，将其用于火灾探测系统中，能够准确、及时地探测各种标准试验火，并具有较强的抗干扰能力．
A structural equivalence fuzzy neural network learning algorithm is proposed. Equivalent neural network according to fuzzy system inference rules, determine the equivalent neural network structure parameters, so the network structure is special. The learning algorithm used is to reverse the weight of the local node to adjust the weights, convergence speed. Experiments show that it can be used in fire detection system to detect all kinds of standard test fire accurately and timely, and has strong anti-interference ability.