复合正交柔性神经网络及其应用

Flexible Neural Network with Compound Orthogonal Type and its Application

针对目前神经网络所存在的不足 ,提出一种带参数的单极性Sigmoid函数的柔性复合正交神经网络 ,并给出相应的参数学习算法 ,这种柔性复合正交神经网络不仅扩大了网络辨识模型的能力与学习适应性 ,而且算法简单 ,学习收敛速度快 ,有线性、非线性逼近精度高等优异特性。以模型辨识作为应用实例 ,仿真结果表明 ,其算法是有效的 。

A kind of flexible neural network of compound orthogonal type with a unipolar sigmoid function was presented,and a parameter and its learning algorithm was given oriented to existing insufficiency of the general compound orthogonal neural network. The flexibly orthogonal neural network not only expands the model identification ability and learning adaptation of the neural network, but also has a simple algorithm, a high-speed convergence of learning process, and excellent characteristics in the linear and nonlinear accurate approximation. The application examples with model identification were given. The simulation results show the learning algorithm in the flexible neural network was effective. Compared with general compound orthogonal neural networks, the performance of this kind of flexible neural network was improved.