摘要
根据Madaline网络工作原理,针对其网络特点和现有算法中存在的缺点,提出将MRⅡ算法与感知机学习规则相结合的改进的学习算法,使每个样本学习后权值的改动尽量小,从而减少它对其它样本学习产生的影响,达到提高网络学习能力的目的。
In this paper, according to the network characteristic and the shortcoming of the existing algorithm, by analyzing the Madaline network work principle and its learning algorithm MRⅡ, propose an improvement algorithm that combines the MRII algorithm and the perception learning rule that will improve the performance of the network learning.
出处
《计算机技术与发展》
2006年第6期224-226,共3页
Computer Technology and Development
