【摘要】 介绍了一种新型的人工神经网络———有监督线性特征映射 (SLFM )网络 ,它综合了BP网络的可监督性和SOM网络算法简单的优点 ,具有学习速度快、精度高、扩展能力较强的优点。文中讨论了SLFM网络的拓扑结构和学习机制 ,并对网络的学习算法进行了改进 ,对比实验表明 ,改进后的SLFM网络其性能得到了进一步的提高。更多还原

【Abstract】 In this paper, a new type of artificial neural network, called supervised linear feature mapping (SLFM) network is presented, which integrates the advantages of BP in which learning is supervised and SOM network in which the algorithm adopted is simple, and features quick learning speed, high learning accuracy and good extension ability. Based on the discussion on the topology and learning algorithm of SLFM, the learning algorithm adopted is improved. Compared with the original, the improved SLFM has better properties.更多还原