Based on the old algorithm, introduce new Linear Feedback Shift Register, new Central Limit Theorem, Maclaurin Equation to generate uniform distribution, gaussian distribution, exponential distribution's random number respectively and analyze them, compares the difference between new algorithm and old algorithm.
在原有算法的基础上，提出了改进型线性反馈移位寄存法、改进型中心极限定理法和麦克劳林展开法，并对其作测试仿真，比较传统算法与改进型算法之间的优缺点。
Let A be the matrix of independent Gaussian random variables centered at A, each of the variances12s, K be the condition number of matrix A.
异常大的矩阵条件数和增长因子是导致矩阵A病态，继而导致解的误差偏大的主要根源。
To design a gaussian random number generator used for channel simulation and coder performance test, high-speed universal random number and correlated stochastic series generators are studied.
为设计一个用于信道仿真和编码器性能测试的高速高斯随机数发生器，研究了适于硬件实现的高速通用的连续随机变量和随机序列产生法。
Well learn Gaussian mixture models from multivariate data the GA-EM algorithm. This algorithm is caballing of selecting the number of components of the model using the minimum description length criterion.
本文将使用GA_EM算法来学习未知支数的多元高斯混合模型，算法中使用MDL准则作为选择最优分支数的信息准则。
In this paper, a penalized minimum matching distance-guided EM algorithm is discussed. Then, under the framework of Greedy EM, an automatic algorithm with high speed and accuracy is proposed to estimate the component number of the Gaussian mixture model. The effectiveness of the proposed algorithm is finally verified by the simulations of univariate and bivariate Gaussian mixture models.
文中研究了一种采用惩罚性最小匹配距离估计分量数的方法，并结合贪婪EM算法框架，提出了一种可以在进行参数估计的同时快速准确地自动估计高斯混合模型分量数的算法，最后通过一元和二元高斯混合模型的仿真实验验证了该算法的有效性。
Experimental results show that the static 12-D Mel-frequency cepstral coefficient feature is the optimal acoustic feature; the optimal number of Gaussian mixture components per state is 1; the optimal number of tied states after model...
实验表明：12维静态M e l频率倒谱系数是最优的语音特征；HMM模型中的状态模型采用单高斯；对于特定说话人的HMM模型，使用分类与衰退树聚类生成的绑定状态模型个数在3 000左右最优。
In this dissertation, many works have been done on the basis of the summarization of the merits of all kinds of meshless methods after a large number of relevant references are read and studied. The basic ideas, principles of meshless methods and their development in the recent years are introduced synoptically. Applications of Element-free Galerkin method in elastic foundation plate are studied deeply and some useful conclusions are obtained as follows:(1) Comparisons between numerical solutions and analytical solutions for several different styles of weight functions, including Gaussian weight function, spline weight function and singular weight function etc, are performed. The results show that Gaussian weight function is the best one.(2) Ranges of relevant parameters for Gaussian weight function are given.(3) Influences of nodal distribution schemes and the number of Gaussian integrated dots on computational results are discussed.
在本文中，作者在认真阅读和研究所搜集到的大量参考文献、综合各种无网格法优点的基础上，首先，对无网格法的基本思想、基本原理、发展现状作了概括性的介绍；其次，对无单元Galerkin法在弹性地基板计算中的应用进行了较为深入的探讨：(1)分别对高斯型、样条型、奇异型等不同形式的权函数进行比较分析，数值计算结果与级数解析解的比较表明，高斯型权函数效果较好：(2)对高斯型权函数进行参数研究，得到地基板计算中相关参数的选取范围；(3)对不同节点分布方案和不同高斯积分对计算结果的影响等作了有益的讨论。
Experimental results show that the static 12-D Mel-frequency cepstral coefficient feature is the optimal acoustic feature; the optimal number of Gaussian mixture components per state is 1; the optimal number of tied states after model clustering by the classification and regreession tree is about 3000 for speaker-dependent tri-phone HMM models.
实验表明：12维静态Mel频率倒谱系数是最优的语音特徵；HMM模型中的状态模型采用单高斯；对於特定说话人的HMM模型，使用分类与衰退树聚类生成的绑定状态模型个数在3000左右最优。
With the guidance of human eyes'perceptual mechanism for color images, the visual information modeling in color image is treated as a multidimensional data clustering problem. Firstly, re-sampling is carried out to get training data points. Then GMM is introduced to model these data points. Through a modified Gibbs sampling algorithm, the parameters of GMM and the number of Gaussian component are obtained automatically.
该算法从人眼对彩色图像的感知机制出发，把对彩色图像中的视觉信息提取过程视为多维数据的分类问题，首先通过重采样抽取训练数据点集，然后引入GMM对彩色图像中的像素分布进行建模，通过一个改进的Gibbs采样彩色图像建模算法取得模型参数，并实现高斯混合元数目的自动确定；最终实现彩色-灰度转换操作。
The exact expressions for the BER of the UWB system are firstly derived using virtual branch technique by considering independent multipath fading on each diversity branch without assumption for equal averaged SNR over the indoor dense multipath fading channel and this BER performance is theoretically analyzed and are compared and evaulated with simulation under the condition of Gaussian distribution assumption for the MAI. Then, with the simulation for impulses having various pulse durations, results show that pulse duration, the number of pulses per symbol, the number of multipath diversity branches selected and their distribution in the time domain have much influence on the BER performance of this UWB radio system.
建立了相应的UWB无线电信号及系统模型，在理想多址信道条件下，对UWB系统的信息传输速率和多址性能进行了定量测量，并对结果进行了理论分析和仿真评估比较，结果表明：数据调制方式以及脉冲的持续宽度均对UWB系统性能有很大影响，证明了本论文所提方案具有可行性，而且，在相同条件下，窄脉冲实现的UWB系统性能均优于宽脉冲。
Finally, the design scheme of Cellular Automata image edge detect algorithm and Gaussian Random Number generator algorithm applied in the system is presented, and an extended application scheme of data acquisition by the system is also provided.
在此基础上，提出了元胞自动机图像边缘检测算法和高斯随机数产生算法在系统中应用的设计方案，并给出了系统应用于数据采集的扩展方案。
For the case of the half-plane diffraction of vortex Gaussian beams there may exist more than one phase singularities or no phase singularity in the diffraction field. The number and position of phase singularities are dependent on the vortex position at the source plane and propagation distance.
对于位相奇点高斯光束的半屏衍射，衍射场有可能出现多个或没有位相奇点，位相奇点的数量和位置与源平面位相奇点的离轴参数和传输距离有关。
The Hofmann's distribution family encompassing the Poisson, the Negative Binomial and the Poisson Inverse Gaussian distributions can be used for the number of claims.
Hofmann分布族可用于复合Poisson过程中索赔次数的研究。
It is shown by a great number of numerical calculation results and physical analyses that the beam quality depends on the obscure ratio of the rectangular aperture, spherical aberration of the lens and the order of the super Gaussian beam.
大量数值计算和物理分析表明，超高斯光束经有方环光阑透镜后的光束质量与方环遮拦比、透镜球差和超高斯光束的阶数等因素有关。
By changing the Fresnel number the focal switch can also be produced under certain conditions. 3. The focal switch of elegant cosh Gaussian beams focused by an astigmatic lens is studied in detail and the influence of parameters on the focal switch are investigated.
详细研究了复宗量双曲余弦高斯光束通过像散透镜聚焦后的焦开关现象，分析了光束参数、像散参数、菲涅尔数及光学系统几何参数对焦开关现象的影响。
Finally we extend the Fresnel number to the region of complex number, and use it to investigate the diffraction problem of a truncated Gaussian beam.
最后，把菲涅耳数推广到复数域，并用于研究截断高斯光束的衍射问题。
We describe a class of closed-form solutions of the paraxial wave equation, the elegant cosh-squared-Gaussian (EchSG) beams whose parameter related with cosh function part is complex number.
本文引入了近轴近似下波动方程的解&复变量双曲余弦平方-高斯光束，该光束与双曲余弦函数项相关的参数为复数。
While for the Gaussian diffracted beams with w 0/R < 1, the axial light intensity exists specifically evolutive law. The number of extrema is reduced step by step and finally disappeared with the initial Gaussian half width diminishing.
而对于w0/R<1的这一类高斯衍射光束，轴上光强存在特定的演化规律：随着初始高斯半宽度的减小，轴上光强极值个数逐步减少直至全部消失。
Besides, the comparison of Gaussian distribution and aggregation number distribution of toroids has been processed.
此外，还以高斯分布对复曲面的聚合数分布进行了模拟与比较。
This paper principally discusses the training problem of Gaussian basis function classifier which can be used for classification. For basis function classifier, how to correctly initialize the number of network hidden nodes and their parameters is very important.
本文主要研究高斯基函数分类器的训练问题，对基函数分类器来说，如何确定网络的初始隐层节点数和隐层节点参数是一个重要问题。