本文提出了非线性对象神经网络建模的广义自组织学习算法,该算法采用多个局部模型进行建模,扩展了Kohonen自组织学习算法中的局部模型划分机制,且多个局部模型的划分兼顾了输入样本的分布和模型匹配特性.仿真结果表明,广义自组织学习算法明显地提高了建模精度和收敛速度. In this paper, a generalized self-organizing learning algorithm for nonlinear object neural network modeling is proposed. The algorithm uses multiple local models to model and extends the local model partitioning mechanism in Kohonen self-organizing learning algorithm, and the partitioning of multiple local models The distribution of input samples and the matching characteristics of the model are simulated.The simulation results show that the generalized self-organizing learning algorithm obviously improves the modeling accuracy and convergence speed.