一种改进的在线最小二乘支持向量机回归算法

针对一般最小二乘支持向量机处理大规模数据集会出现训练速度幔、计算量大、不易在线训练的缺点,将修正后的遗忘因子矩形窗方法与支持向量机相结合,提出一种基于改进的遗忘因子矩形窗算法的在线最小二乘支持向量机回归算法,既突出了当前窗口数据的作用,又考虑了历史数据的影响.所提出的算法可减少计算量,提高在线辨识精度.仿真算例表明了该方法的有效性.     

电离层延迟模型改进研究

由于电离层延迟会对卫星的定位精度产生很大的影响,传统的电离层延迟Klobuchar模型的预报精度不高,而且其对夜间电离层延迟以及对延迟峰值出现时刻的预报存在明显的不合理性。因此必须进行修正,根据实测电离层延迟数据的特点对传统的Klobuchar模型进行了改进,并利用非线性最小二乘迭代算法对两种模型进行参数拟合,采用拟合参数计算两种模型的拟合值。通过比较两种拟合值相对于实际测量值的精度,可以得出:改进模型比传统模型提高了电离层延迟预报的精度,并且改善了传统延迟模型预报中存在的不合理性。     

Effect of pressure on three catalytic partial oxidation reactions at millisecond contact times

The effects of pressure on reactant conversion and product selectivities in three catalytic oxidation systems have been examined at pressures between 1 and > 5 atm. Reaction was sustained autothermally near adiabatic operating conditions at temperatures of 1000°C with residence times over the noble metal catalysts between 10^–4 and 10^–2 s. The three systems investigated were (1) HCN synthesis over Pt-10% Rh gauze catalysts, (2) methane oxidation to synthesis gas (CO and H₂) over rhodium-coated monoliths, and (3) ethane conversion to ethylene over platinum-coated monoliths. We find that selectivities in all three reactions do not change dramatically with approximately a five-fold increase in pressure. This strongly suggests that free radical homogeneous chain reactions are not significant in these processes and that they can be operated reliably above atmospheric pressure. For the synthesis of HCN over Pt-10% Rh gauzes, the selectivity to HCN can be maintained above 0.75 at pressures up to 5.5 atm. Selectivities to synthesis gas (CO and H₂) from a methane-air mixture over a Rh-coated foam monolith at pressures up to 5.5 atm were maintained above 0.90. Over a Pt-coated foam monolith, the selectivity to ethylene from ethane-air and ethane-O₂ mixtures was independent of pressure up to 6.5 atm and conversion rose slightly although it was necessary to maintain constant velocity and residence time over the catalyst to avoid carbon formation.This research was supported by DOE under Grant No. DE-FG02-88ER13878.     

Morphological Models for the Collapse of Area Features in Digital Map Generalization

Collapse is an essential operation for the manipulation of area features in digital data generalization. This operation can be categorized into two types: complete collapse and partial collapse. The former is composed of another two types: area-to-point and area-to-line collapse. In this paper, a set of algebraic models built upon the operators in mathematical morphology is described for the area-to-line collapse and partial collapse operations. For the area-to-line collapse operation, a modified skeleton algorithm is presented. For the partial collapse operation, a procedure is designed which consists of a set of operations, i.e., the skeletonization, separation of areal and linear parts, simplification of areas and an overlay operation. These models are tested using real map data sets.     

基于联合迭代重建技术的超声CT重建算法

在研究超声CT成像原理的基础上,对阵列检测方法进行了探讨。通过对探头布置方式的改进,增加了成像所需的数据量。基于最小二乘方准则,利用联合迭代重建算法,采用四边扫描算法求解相关系数矩阵,通过不断修正误差逼近真实值,进而实现速度矩阵的求取。数值仿真表明：采用上述方法能够明显减小重建误差,改善层析成像效果。     

齐次等式约束线性回归模型条件根方估计的效率

对齐次等式约束线性回归模型回归系数的约束最小二乘估计提出改进,引入一种估计的相对效率,证明在一定条件下,狭义条件根方估计、广义条件根方估计的效率均高于约束最小二乘估计的效率．     

Geoadditive expectile regression

Quantile regression has emerged as one of the standard tools for regression analysis that enables a proper assessment of the complete conditional distribution of responses even in the presence of heteroscedastic errors. Quantile regression estimates are obtained by minimising an asymmetrically weighted sum of absolute deviations from the regression line, a decision theoretic formulation of the estimation problem that avoids a full specification of the error term distribution. Recent advances in mean regression have concentrated on making the regression structure more flexible by including nonlinear effects of continuous covariates, random effects or spatial effects. These extensions often rely on penalised least squares or penalised likelihood estimation with quadratic penalties and may therefore be difficult to combine with the linear programming approaches often considered in quantile regression. As a consequence, geoadditive expectile regression based on minimising an asymmetrically weighted sum of squared residuals is introduced. Different estimation procedures are presented including least asymmetrically weighted squares, boosting and restricted expectile regression. The properties of these procedures are investigated in a simulation study and an analysis on rental fees in Munich is provided where the geoadditive specification allows for an analysis of nonlinear effects of the size of flats or the year of construction and the spatial distribution of rents simultaneously.     

Least squares support vector machines with tuning based on chaotic differential evolution approach applied to the identification of a thermal process

In the past decade, support vector machines (SVMs) have gained the attention of many researchers. SVMs are non-parametric supervised learning schemes that rely on statistical learning theory which enables learning machines to generalize well to unseen data. SVMs refer to kernel-based methods that have been introduced as a robust approach to classification and regression problems, lately has handled nonlinear identification problems, the so called support vector regression. In SVMs designs for nonlinear identification, a nonlinear model is represented by an expansion in terms of nonlinear mappings of the model input. The nonlinear mappings define a feature space, which may have infinite dimension. In this context, a relevant identification approach is the least squares support vector machines (LS-SVMs). Compared to the other identification method, LS-SVMs possess prominent advantages: its generalization performance (i.e. error rates on test sets) either matches or is significantly better than that of the competing methods, and more importantly, the performance does not depend on the dimensionality of the input data. Consider a constrained optimization problem of quadratic programing with a regularized cost function, the training process of LS-SVM involves the selection of kernel parameters and the regularization parameter of the objective function. A good choice of these parameters is crucial for the performance of the estimator. In this paper, the LS-SVMs design proposed is the combination of LS-SVM and a new chaotic differential evolution optimization approach based on Ikeda map (CDEK). The CDEK is adopted in tuning of regularization parameter and the radial basis function bandwith. Simulations using LS-SVMs on NARX (Nonlinear AutoRegressive with eXogenous inputs) for the identification of a thermal process show the effectiveness and practicality of the proposed CDEK algorithm when compared with the classical DE approach.     

甲烷部分氧化制合成气的研究进展

综述了近年来甲烷部分氧化制合成气的催化剂，反应机理及活性中心的研究进展及反应中存在的主要问题。     

Meshless local Petrov-Galerkin method for continuously nonhomogeneous linear viscoelastic solids

A meshless method based on the local Petrov-Galerkin approach is proposed for the solution of quasi-static and transient dynamic problems in two-dimensional (2-D) nonhomogeneous linear viscoelastic media. A unit step function is used as the test functions in the local weak form. It is leading to local boundary integral equations (LBIEs) involving only a domain-integral in the case of transient dynamic problems. The correspondence principle is applied to such nonhomogeneous linear viscoelastic solids where relaxation moduli are separable in space and time variables. Then, the LBIEs are formulated for the Laplace-transformed viscoelastic problem. The analyzed domain is covered by small subdomains with a simple geometry such as circles in 2-D problems. The moving least squares (MLS) method is used for approximation of physical quantities in LBIEs.