最小二乘分解算法在车型识别中的应用

提出一种最小二乘支持向量机的序贯最小分类分解算法.针对最小二乘支持向量机,通过对核函数的相关变换,将二阶的误差信息归结到优化方程的一阶信息中,从而简化运算过程.采用最优函数梯度二阶信息选择工作集,实现最小二乘支持向量机分解算法,提高了算法的收敛性.采用径向基核函数和交叉验证网格搜索的方法验证算法的分类准确性.实验结果表明,提出的分类算法应用于车型识别中,可以得到比其他分类方法更好的分类准确度.     

混合核函数稀疏LS-SVM软测量建模与应用

针对最小二乘支持向量机存在的稀疏性欠缺和单核函数局限性问题,本文提出一种基于混合核函数稀疏最小二乘支持向量机的软测量建模方法.该方法使用多项式核函数和RBF核函数线性加权构成混合核函数,兼顾最小二乘支持向量机的全局拟合能力与局部拟合能力,以矢量基学习作为稀疏解算法,改善最小二乘支持向量机的稀疏性,在精简模型结构的同时,避免冗余信息中的噪声过多的拟合到模型参数中,进而采用粒子群算法优化模型部分参数.将此方法分别应用于Mackey- Glasss混沌模型的时间序列预测和乙烯精馏塔塔釜乙烯浓度预测,应用结果表明该方法较最小二乘支持向量机、稀疏最小二乘支持向量机以及混合核最小二乘支持向量机具有更好的泛化效果和预报精度,兆示出其良好的应用潜力.     

基于最小二乘支持向量机的函数拟合的研究

函数拟合通常要在有限的训练样本下对函数变量之间的关系做出预测,在实践中由于训练样本有限,并且训练样本本身存在噪音和孤立点,用传统的方法进行函数拟合的结果往往不能满足要求.本文主要利用最小二乘支持向量机对函数进行拟合.首先介绍了最小二乘支持向量机的工作原理,并对参数选择方法进行了研究,然后通过仿真实验对利用最小二乘支持向量机进行函数拟合的效果加以对比说明.     

一种井下人员无线定位算法研究

针对传统井下指纹定位算法存在需要采集大量指纹数据和定位精度不高的问题,提出了一种差分鱼群优化最小二乘支持向量机(DEAFSA-LSSVM)的井下人员无线定位算法.首先将井下实验区域划分为多个小区域,并利用克里金插值算法建立指纹数据库;然后利用差分进化与人工鱼群混合智能算法优化正则化参数和核函数宽度,建立最小二乘支持向量机算法模型,利用无线采集接收终端采集待定位点的无线信息数据,通过最小二乘支持向量机算法模型计算出其所属小区域;最后利用小区域内无线信息数据,通过加权K近邻算法进行实时定位.实验结果表明:该定位算法的收敛速度快,分类准确,准确率达到98.87%;定位精度高,平均定位误差为1.51 m,比未经优化的最小二乘支持向量机算法的定位精度提高18.82%.     

基于EEMD-GWO-LSSVM的公共交通短期客流预测

为了提高大型公共交通短期客流预测精度,提出了一种在利用集成经验模态分解原始数据的条件下,采用灰狼优化算法优化最小二乘支持向量机（EEMD-GWO-LSSVM）的算法,利用该算法实现城市大型公共交通短期客流预测。该模型采用EEMD分解原始数据,将分解后的各个本征模函数（IMF）分量运用最小二乘支持向量机进行回归预测,最小二乘支持向量机的预测参数由灰狼算法进行优化。通过对西安地铁二号线北客站一个月进出站人数进行训练预测,将预测结果和支持向量机（SVM）,自回归移动平均模型（ARIMA）,仅利用灰狼优化参数的最小二乘支持向量机（GWO-LSSVM）算法以及基于交叉检验进行参数优化的最小二乘支持向量机进行对比,分析得出该算法具有更加精确的预测结果。     

基于改进PSO优化LSSVM的传感器补偿研究

针对最小二乘支持向量机在对传感器进行补偿时,正则化参数和核函数参数对补偿精度影响较大的问题,提出一种利用改进的粒子群优化算法优化最小二乘支持向量机模型参数的传感器补偿方法。该方法利用改进的粒子群优化算法优化最小二乘支持向量机模型的正则化参数和核函数参数,避免了人工选择参数的盲目性,提高了最小二乘支持向量机模型的预测精度。仿真实验表明,在传感器的补偿时,该方法比最小二乘支持向量机模型的补偿精度更高。     

离散数据拟合模型的研究与实现

最小二乘支持向量机引入到离散数据拟合中,代替传统的最小二乘法解决离散数据拟合问题。推导了用于函数估计的最小二乘支持向量机算法,构建了基于最小二乘支持向量机的离散数据拟合模型,并对电机数据拟合进行了研究。结果表明,最小二乘支持向量机拟合离散数据比最小二乘法精度更高、拟合效果更好。     

变异粒子群优化最小二乘支持向量机的网络流量预测

针对最小二乘支持向量机参数优化问题,提出一种变异粒子群算法优化最小二乘支持向量的网络流量预测模型(MPSOLSSVM)。首先对网络流量序列进行相空间重构,构建最小二乘支持向量的学习样本;然后采用变异粒子群算法选择最小二乘支持向量机参数,从而建立最优的网络流量预测模型,最后与其他模型进行对比实验。对比结果表明,相对于对比模型,MPSO-LSSVM提高了网络流量的预测精度,预测结果可以为网络管理员提供有价值参考信息。     

最小二乘隐空间支持向量机

在隐空间中采用最小二乘损失函数,提出了最小二乘隐空间支持向量机（LSHSSVMs）．同隐空间支持向量机（HSSVMs）一样,最小二乘隐空间支持向量机不需要核函数满足正定条件,从而扩展了支持向量机核函数的选择范围．由于采用了最小二乘损失函数,最小二乘隐空问支持向量机产生的优化问题为无约束凸二次规划,这比隐空间支持向量机产生的约束凸二次规划更易求解．仿真实验结果表明所提算法在计算时间和推广能力上较隐空间支持向量机存在一定的优势．     

基于PSO的LS-SVM特征选择与参数优化算法

针对最小二乘支持向量机特征选择及参数优化问题,提出了一种基于PSO的LS-SVM特征选择与参数同步优化算法。首先产生若干种群（特征子集）,然后用PSO算法对特征及参数进行优化。在UCI标准数据集上进行的仿真实验表明,该算法可有效地找出合适的特征子集及LS-SVM参数,且与基于遗传算法的最小二乘支持向量机算法（GALS-SVM）和传统的LS-SVM算法相比具有较好的分类效果。     

核聚类人工免疫网络的参数研究

核聚类人工免疫网络是一种新型的聚类分析方法,其基本算法形式也是启发式随机搜索算法,所以算法中的参数对于方法的最后性能有着重要影响。本文通过实验探讨几个关键参数对算法性能的影响,从而说明算法中参数的选择方式。     

The use of the generalized links to determine the minimum inertial parameters of robots

This article presents a new method to determine the minimum set of inertial parameters of robots. The method permits to determine most of the regrouped parameters by means of recursive closed form relations. It is based on the use of a new expression for the kinetic and potential energy of the robot as function of the inertial parameters of the generalized links. The demonstration is seen to be easier than the previous methods and gives a clear idea about the physical meaning of the regrouped parameters.     

同步挤压小波变换提取地震信号瞬时属性

地震信号瞬时参数的准确提取对于研究地球介质的岩性构造、油气储藏等具有重要意义。针对用传统时频分析方法提取地震信号瞬时参数对噪声敏感、可靠性不强等问题,提出了利用同步挤压小波变换(SST)提取地震信号瞬时参数的方法。该方法是一种连续可逆的时频变换工具,首先结合时频谱重排的思想,得到同步挤压小波变换量值,然后计算出实信号对应的解析信号,接着求取瞬时参数,最后与传统的Hilbert变换方法进行比较,通过仿真信号试验和实际地震资料的应用表明,SST提取的地震瞬时参数抗噪性更强、可靠性强,尤其对于瞬时频率的提取。     

人工目标图像中的消失点检测

在人工目标检测中,消失点信息比边缘、拐角等特征更加鲁邦。根据累计空间的不同,消失点检测方法可划分为有限空间检测和无限空间检测两大类。在有限空间检测方法中,Barnard提出的基于高斯映射的消失点检测,因同时适用于有限远和无穷远消失点的检测,最为引人关注。但该方法因累计空间划分不均匀,容易造成统计计算的偏差。文章针对人工目标固有特征,提出了长度优先的向量叉积候选消失点检测方法,在避免了因累计单元大小不均而造成计算偏差的同时,强调了人工目标的主体特征。同时利用三个彼此正交的方向信息,反算相机的内部参数,用以判定三个正交方向选择的合理性。从而使得消失点的确定,能够更好地满足人工目标特征的提取要求。     

用MATLAB仿真分析自抗扰控制器的整定参数

自抗扰控制器在实际使用时有多个需要整定的参数,由于控制器参数多,又无法确定参数作用方向,因此参数整定困难,为实际使用自抗扰控制器带来不便。本文利用MATLAB软件仿真分析控制器及参数,寻求参数整定规律,简化整定过程,对实际应用自抗扰控制器具有较大的意义。     

Dynamic analysis of structures with uncertain parameters using the transfer matrix method

A combination of the transfer matrix method and a first order, second moment approach is used to perform dynamic analyses of structures with uncertain parameters. This technique provides relations between the vibrational characteristics of structures with uncertain parameters and the first and second statistical moments of these parameters. The proposed methodology leads to a straightforward analysis with calculations that are much simpler than those with other methods, such as the stochastic finite element method. The effectiveness of this method is verified by a numerical example.     

Estimation of the parameters in regularized simultaneous super-resolution

We describe a method for automatic determination of the regularization parameters for the class of simultaneous super-resolution (SR) algorithms. This method, proposed in (Zibetti et al., 2008c), is based on the joint maximum a posteriori (JMAP) estimation technique, which is a fast alternative to estimate the parameters. However, the classical JMAP technique can be unstable and may generate multiple local minima. In order to stabilize the JMAP estimation, while achieving a cost function with a unique global solution, we derive an improved solution by modeling the JMAP hyperparameters with a gamma prior distribution. In this work, experimental results are provided to illustrate the effectiveness of the proposed method for automatic determination of the regularization parameters for the simultaneous SR. Moreover, we contrast the proposed method to a reference method with known fixed parameters as well as to other parameter selection methods based on the L-curve. These results validate the proposed method as a very attractive alternative for estimating the regularization parameters.     

An effective approach for parameter determination of the digital phase-locked loop in the z-domain

In digital communication systems, typical methodologies in determining loop parameters of the digital phase-locked loop (DPLL) are based on the mapping transformation from the analog domain to the digital domain. However, such transform based algorithms are relatively complicated and not straightforward, and they also cause the problem that loop parameters are affected by the pre-detection integration time greatly. To solve these issues, an effective direct method of determining loop parameters of the second-order DPLL in the z-domain is proposed in this paper. Through ascertaining specific positions of the closed-loop system function's poles inside the right-hand side of the z-plane's unit circle, unknown parameters are calculated directly and flexibly in this method, which enables the DPLL to acquire good low-pass filtering characteristic and system stability. This novel method not only reduces the complexity of solving the parameters, but also eliminates the effect of the pre-detection integration time on loop parameters. Simulation results are provided to confirm the feasibility of the proposed method and to show that the DPLL obtained by this method achieves the similar tracking performance to the discretized PLL.     

A method to detect and characterize ellipses using the Houghtransform

We describe a new technique for detecting and characterizing ellipsoidal shapes automatically from any type of image. This technique is a single pass algorithm which can extract any group of ellipse parameters or characteristics which can be computed from those parameters without having to detect all five parameters for each ellipsoidal shape. Moreover, the method can explicitly incorporate any a priori knowledge the user may have concerning ellipse parameters. The method is based on techniques from projective geometry and on the Hough transform. This technique can significantly reduce interpretation and computation time by automatically extracting only those features or geometric parameters of interest from images and making exact use of a priori information     

Choice of B-splines with free parameters in the flexible discriminant analysis context

Flexible discriminant analysis (FDA) is a general methodology which aims at providing tools for multigroup non linear classification. It consists in a nonparametric version of discriminant analysis by replacing linear regression by any nonparametric regression method. A new option for FDA, consisting in a nonparametric regression method based on B-spline functions, will be introduced. The relevance of the transformation (hence the discrimination) depends on the parameters defining the spline functions: degree, number and location of the knots for each continuous variable. This method called FDA-FKBS (Free Knot B-Splines) allows to determine all these parameters without the necessity of many prior parameters. It is inspired by Reversible Jumps Monte Carlo Markov Chains but the objective function is different and the Bayesian aspect is put aside.