摘要
介绍利用自组织神经网络模型及算法对已知样本和未知样本进行聚类 ,然后将各聚类中的样本分别提交给多个BP子网处理来降低BP网络学习复杂度。
How to combine the Self-Organizing feature Map Model with the Back Error Propagation neural network to reduce the learning complexity of the BP Algorithm is introduced in the paper.The samples for BP can be aggregated with the SOM model.Each group of aggregated samples is trained with a Bp model.
出处
《新疆石油学院学报》
2001年第3期77-80,共4页
Journal of Xinjiang Petroleum Institute
