An improved neural network based on L-M algorithm has been applied to fault diagnosis expert system against to the slow convergence rate of conventional BP neural network.
针对传统BP神经网络训练中收敛速度较慢的缺点,提出一种基于L - M算法的神经网络应用于机械设备故障诊断的专家系统。
With L-M regular training it can promote the speed and precision of net convergence.
用l - M规则训练可使网络收敛加快精度提高。
Theoretical analysis proves that the M-PSO algorithm keeps convergence.
通过理论分析,证明算法具有良好的收敛性。
The network could predict the effects of the temperature on the boundary film strength. the network trained with the L-M rule were quick in convergence and small in error.
该模型可用于准确地预测温度对边界膜强度的影响。并采用L - M规则进行神经网络学习训练可使网络收敛快,误差小。
It was concluded that M-mode and two-dimensional color Doppler flow convergence region could be used to quantificate the area of mitral stenosis accurately.
结果表明,M型彩色多普勒血流会聚法与二维彩色多普勒血流会聚法在检测二尖瓣狭窄瓣口面积方面基本可靠,但两者的准确性无显著差异。
By introduction of the ideology of population migration, the M-PSO algorithm keeps the convergence and has good performance such as optimization velocity and optimization results.
通过引入人口迁移的思想,在保证算法收敛性的同时,使M - PSO算法具有良好的优化速度和优化效果。
Phoria at distance and near, positive and negative relative accommodation(NRA/PRA), accommodative convergence and accommodation response were measured. Results Response AC/A ratio of P. M.
眼动参数测量包括远、近距隐斜、调节反应、调节性辐辏及正、负相对调节,对每组的反应性、计算性AC/A 等眼动参数进行分析比较。
Phoria at distance and near, positive and negative relative accommodation(NRA/PRA), accommodative convergence and accommodation response were measured. Results Response AC/A ratio of P. M.
眼动参数测量包括远、近距隐斜、调节反应、调节性辐辏及正、负相对调节,对每组的反应性、计算性AC/A 等眼动参数进行分析比较。
应用推荐