论文部分内容阅读
根据对Hopfield神经网络改进模型大量的计算模拟实验,证明了这种存储矩阵对角元T_(ii)=1的改进模型具有比T_(ii)=0的原模型更大的存储容量和更强的纠错能力,并通过对二元矢量的编码,使得这个模型存储容量和纠错能力得到进一步提高。
According to a large number of computational experiments on the improved Hopfield neural network model, it is proved that the improved model with T_ (ii) = 1 for the storage matrix has more storage capacity and stronger performance than the original model with T_ (ii) = 0 The error correction capability of the model is further improved by coding the binary vector.