提出了自适应学习率及动量因子的BP神经网络算法和误差逼近度渐近收缩学习的BP神经网络算法，并将其应用于汽轮发电机组振动故障诊断与识别。
The improved BP algorithms based on adaptive parameters adjustment and error contracting gradually are presented, which are applied successfully to fault diagnosis of steam- turbine generator unit.
沪指开盘后震荡上攻，盘中一度逼近3100点大关。
The Shanghai Composite Index jumped after opening, but despite suffering some fluctuations reached near 3100 points during trading.
随着计算能力的增加，以及传感器、效应器与显示器的改进，虚拟现实将开始在保真度和互动性上逼近物理现实。
As computational power increases, and as sensors, effectors and displays improve, VR could begin to approximate physical reality in terms of fidelity and interactivity.
提出球度误差评定的几何逼近原理，并对算法进行了建模。
The geometrical approximation theory of evaluating geometrical errors is put forward and modeled.
提出了一种新的结合特征灵敏度直接法和向量值函数有理逼近的结构动力重分析方法。
The new method is based on combining perturbation method of eigen-problems with rational approximation of a vector-value function .
采用改进型逐次逼近解法(MSAM)计算了积分方程，该方法所需内存量小，且可适用于高电导率对比度地层。
The integral equations were computed by modified successive approximation method (MSAM). The method needs less computer memory requirements and is suitable for high-conductivity-contrast formation.
文中用逼近法和双线性变换法，设计了用于圆度测量的高斯数字逼近滤波器，并给出了零相移的递归滤波算法，计算量小，计算效率高，易于实现。
A series of Gaussian digital approximation filters used in roundness measurement were designed on the basis of approximation method and bilinear transformation.
曲面轮廓度误差评价和多边域逼近均要求对理论曲面进行数字化自适应采样。
Adaptive sampling algorithm of surface play an important role in evaluation of profile deviation and surface triangular approximation.
本文讨论了二阶障碍问题的一个五自由度非协调矩形元逼近。
In this paper, a five-degree nonconforming rectangle element approximation for second order obstacle problem is considered.
提出用逐次逼近思想来评定复杂曲线的轮廓度误差的方法。
An approaching method was proposed to evaluate the profile error of curve.
本文提出沿圆柱表面的螺旋线对圆柱度作动态测量，并应用最佳一致逼近方法对其展开的测量数据来评定圆柱度误差值。
This paper presents a method to evaluate cylindricity errors. It USES best uniform approximation to process the sample data measured continually and dynamically along a cylindrical spiral.
针对不同重要性的样本，采用不同的惩罚因子进行逼近，在训练错误率和模型复杂度之间进行权衡。
Different penalty factors are assigned to the sample data of different importance, and the tradeoff can be determined between training errors and model complexity.
应用最佳一致逼近理论，从最小条件出发建立了评定平面度误差的数学模型，对评定平面度误差的理论问题进行了分析研究。
In this paper, the mathematical model of flatness error is established. The theory of optimal approximation is used to analyze the theoretical problem of flatness error.
跨越这一障碍的有效方法之一是采用参数矩阵的低秩逼近，目的是控制模型复杂度。
Low rank approximation to parametric matrix has recently been proven to be an effective method to control the complexity of models.
采用埃尔米特插值多项式代替泰勒展开式逼近目标函数，极大地降低了高阶交叉灵敏度的阶数。
Adopting the Hermite polynomial instead of the Taylor polynomial can greatly decrease the order of high-order cross sensitivity.
影响检测精度的因素包括：视感度滤波器的逼近；
The factors that affect detection precision include: visual acuity filter;
影响检测精度的因素包括：视感度滤波器的逼近；
The factors that affect detection precision include: visual acuity filter;
应用推荐