摘要
作者用目前广泛使用的ＢＰ网络和不太被人们熟悉的ＣＭＡＣ网络进行仿真实验，通过解决三个不同类型的问题（心电图分类、坐标变换、预测），从几个方向比较了这两个原理上都是非线性映射、但在实现上却有很大不同的网络的功能和特性，以期对理解、选择及改进神经网络模型和算法有所帮助。
ome emulation tests were made on BP network, which is widely used and CMACnetwork not well known. A comparison of function and character was made between thesetransformation, prediction). The two networks were all classified as nonliner mapping, buthad quite difference in realizing. We hope this emulation comparison will be helpful forpeople in understanding, choosing and improving the neural networks and algorithms.
出处
《北京生物医学工程》
1997年第2期115-121,共7页
Beijing Biomedical Engineering
