2) weight initialization
权值初始化
1.
A novel learning algorithm is proposed that is based on the combination of independent component analysis(ICA)based weight initialization and automatically adjusting the gain parameter of sigmoid activation function.
提出了一种基于独立元分析(ICA)方法的权值初始化方法和动态调整S型激励函数的斜率相结合的神经网络学习算法。
3) network initialization
网络初始化
1.
Aiming at the disadvantage of traditional BP network,the stability and generalization performance of inverted efferent nerve network in the forecast of optimal asphalt content are increased according to improvement of network initialization,network structure and the use of dynamic network learning mechanism.
利用反传神经网络算法由沥青混合料级配等参数作为输入参数估计沥青混凝土最佳沥青含量,针对传统BP网络的缺陷,通过改进网络初始化、网络结构,采用动态的网络学习机制,提高了反传神经网络在最佳沥青含量预估中的稳定性和泛化性能。
4) weights of artificial neural networks
网络初始权重
5) weight optimal initialization
初始权值优化
1.
Furthermore, a weight optimal initialization method is introduced for improving performance of the soft-sensing model.
针对回归神经网络训练效率低,泛化能力差等问题,尝试引入一种初始权值优化方法加以改进。
6) weights initialization
初始权值
1.
The weights initialization of input floor and output floor are set by applying the loading weights of dependent variable and cause variable,the member of hidden nodes are set by applying factor numbers o.
采用非线性迭代偏最小二乘算法预处理数据,将得到主成分数、自变量和因变量的主成分数的权重以及主成分间的关系矩阵B,以此用来确定BP网络的隐节点数和输入层、输出层的初始权值以及隐节点的关联系数。
2.
The weights initialization and the member of hidden nodes are set by applying theloading weights and factor numbers of O-PLS algorithm.
介绍一种新的多变量数据预处理方法——正交信号修正(OSC)法,提出一种OSC与NIPALS算法结合的O-PLS算法,将该方法用于确定BP网络的基本结构,即确定BP网络的隐层数、节点数及其初始权值,由此建立了O-PLS-BP网络模型。
补充资料：因侵害姓名权、肖像权、名誉权、荣誉权产生的索赔权
因侵害姓名权、肖像权、名誉权、荣誉权产生的索赔权：公民、法人的姓名权、名称权，名誉权、荣誉权、受到侵害的有权要求停止侵害，恢复名誉，消除影响，赔礼道歉，并可以要求赔偿损失。
说明：补充资料仅用于学习参考，请勿用于其它任何用途。
参考词条