基于模糊算法的判决神经网络

Fuzzy Based DecisionBased Neural Network

在判决神经网络（ＤＢＮＮ）的基础上提出了一种基于模糊算法的模糊判决神经网络（ＦＤＢＮＮ）．在网络训练中引入置信度和容噪度的概念，提高了网络分类的稳定性，同时克服了（ＤＢＮＮ）在训练样本混有噪声时学习困难和泛化能力不高的缺点．因ＦＤＢＮＮ在学习时的不均匀性，大大加快了网络训练的时间，提高了训练的效率．实验结果表明，ＦＤＢＮＮ的性能高于ＢＰ网，而且也比ＤＢＮＮ在稳定性和识别率上有了显著的提高．

A fuzzy based decisionbased neural network(FDBNN) based on DBNN(decisionbased neural networks) is proposed in this paper, which introduces conceptions of vigilance and tolerance.FDBNN not only increases the stability of classification, but also improves the performance of learning and ability of generalization when training patterns contain noises. The nonregularity of training saves the learning time enormously and improves the efficiency of training. The experimental results show that the performance of FDBNN is better than that of BP and DBNN on stability and recognition rate.