摘要
首先分析了Hopfield神经网络的存储容量,给出了单个神经元输出在联想过程中的出错概率Pc一定时存储容量的典型数据,分析了整个模式出错概率Pc-定时,最大模式数与网络规模的关系,并用切比雪夫方法对分析结果进行了拟合。
The memory capacity of Hopfield Neural Networks is analyzed and the typical data of the memory capacity is given on condition that the error probability Pc of a single neuron's output is constant. And also when the whole pattern's error probability Pc is con- stant, the relation equation between the capacity and the network scale is given. And the re- sults is fitted by using the Chebyshev method.
出处
《指挥技术学院学报》
1995年第2期78-82,共5页
Journal of Institute of Command and Technology