1983年，Sejnowski和Hinton提出了大规模并行网络学习机即后来的Boltzmann学习机，并明确提出隐单元（Hidden Unit）的概念。
基于48个网页-相关网页
A novel method using entropy to estimate the number of hidden units was proposed. The simulation outcome showed that the effect of BP model with the number of hidden units generated by entropy performed better than other model.
通过数据模拟与现有几种隐单元数确定方法的效果进行比较，熵法构造的模型的判别效果优于其它隐单元数模型。
参考来源 - BP神经网络构建与优化的研究及其在医学统计中的应用·2,447,543篇论文数据，部分数据来源于NoteExpress
结果表明，适当的改进动量因子和隐单元个数将加快网络的收敛性，提高预测能力。
The result presents that improve properly momentum gene and the number of concealed cells will quicken the astringency of network and increase ability of forecast.
本文改进了递归正交最小二乘（ROLS）算法的停止条件，并用改进的ROLS算法优选RBF神经网络中隐单元的个数；
In this paper, the stop condition for recursion orthogonal least square (ROLS) algorithm is improved, and the optimal number of hidden neurons in RBFNN is chosen using this improved ROLS algorithm.
在对面板坝进行脱空分析时，主要采用隐单元的方法模拟现实中出现的脱空，为了满足计算的精度，在要脱空的区域进行了局部加密。
Void was simulated with deleting element as analyzing void of concrete face, and the grid of meshes were increased at requiring void area for meeting the computational accuracy.
应用推荐