摘要
通过对Elman网络的研究 ,提出了一种新型的基于输入层、隐层、输出层神经元递归的动态递归神经网络 ,并给出了其算法。通过在系统辨识中的应用表明 ,该网络收敛速度快 ,模型精度高 ,具有较为广阔的应用前景。
Through study on Elman network, this paper provide a new neural dynamic recurrent neural network based on input, output, hide layer and give its algorithms. By the use of it in system identification, we can conclude that this network could convergent quickly and provide high precision of models. This network have bright outlook in automatic application field.
出处
《计算机仿真》
CSCD
2002年第5期64-66,共3页
Computer Simulation
