摘要
针对已有的三层感知器工作原理,通过增加多个隐层的方法,将其扩展成为基于BP算法的多层感知器网络;利用单、双极性Sigmoid函数的性质和多元函数的偏导数求导法则,推导出了多层感知器网络工作原理,并且给出BP算法的编程步骤;最后的实例证明了本文算法的有效性.
A multilayer perceptron network is extended to based on BP algorithm,according to the working principle of the three layer perceptron,through increasing the number of hidden layer,using the law of partial derivative derivative properties of single,and bipolar Sigmoid function and multiple function,deduced the principle of multilayer perceptron network,and are described in forms of flow chart based on the multi-layer perceptron neural network BP algorithm program.
出处
《嘉应学院学报》
2015年第8期18-21,共4页
Journal of Jiaying University
