A tutorial on support vector machine-based methods for classification problems in chemometrics

This tutorial provides a concise overview of support vector machines and different closely related techniques for pattern classification. The tutorial starts with the formulation of support vector machines for classification. The method of least squares support vector machines is explained. Approaches to retrieve a probabilistic interpretation are covered and it is explained how the binary classification techniques can be extended to multi-class methods. Kernel logistic regression, which is closely related to iteratively weighted least squares support vector machines, is discussed. Different practical aspects of these methods are addressed: the issue of feature selection, parameter tuning, unbalanced data sets, model evaluation and statistical comparison. The different concepts are illustrated on three real-life applications in the field of metabolomics, genetics and proteomics.     

Implicit Shape Reconstruction of Unorganized Points Using PDE-Based Deformable 3D Manifolds

<正>In this work we consider the problem of shape reconstruction from an unorganized data set which has many important applications in medical imaging,scientific computing,reverse engineering and geometric modelling.The reconstructed surface is obtained by continuously deforming an initial surface following the Partial Differential Equation(PDE)-based diffusion model derived by a minimal volume-like variational formulation.The evolution is driven both by the distance from the data set and by the curvature analytically computed by it.The distance function is computed by implicit local interpolants defined in terms of radial basis functions.Space discretization of the PDE model is obtained by finite co-volume schemes and semi-implicit approach is used in time/scale.The use of a level set method for the numerical computation of the surface reconstruction allows us to handle complex geometry and even changing topology, without the need of user-interaction.Numerical examples demonstrate the ability of the proposed method to produce high quality reconstructions.Moreover,we show the effectiveness of the new approach to solve hole filling problems and Boolean operations between different data sets.     

Intelligent active vibration control in an isolation platform

This paper presents the study on the vibration attenuation in an isolated platform by combining multi-layer perception (MLP) neural network, radial basis function (RBF) neural network, cerebella model articulation controller (CMAC) neural network and fuzzy neural networks (FNN) with error back propagation algorithm to control voice coil actuator. Usually, the methods in past time to control vibration were mainly designed by using mathematical models, which must be nearly close to the actual plant models. As regards to these utilized control methods, the most important advantage of them are that they have capability of self tuning the parameters of controllers and could adapt the changes of the environments. The performance of attenuation and control effectiveness can be evaluated by placing the accelerator to measure the amplitude at the center of the isolated platform. The experimental results in this study show that the control methods as adopted could greatly attenuate the vibration of resonance and external disturbance in an isolation platform.     

基于模糊与神经网络技术的复杂非线性跟踪控制

针对一类具有不确定性、多重时延和状态未知的复杂非线性系统,把模糊T-S模型和RBF神经网络结合起来,提出了一种基于观测器的跟踪控制方案.首先,应用模糊T-S模型对非线性系统建模,设计观测器用来观测系统状态,并由线性矩阵不等式得到模糊模型的控制律;其次,构建了自适应RBF神经网络,应用自适应RBF神经网络作为补偿器来补偿建模误差和不确定非线性部分.证明了闭环系统满足期望的跟踪性能.示例仿真结果表明了该方案的有效性.     

RBF collocation method and stability analysis for phononic crystals

The mesh-free radial basis function (RBF) collocation method is explored to calculate band structures of periodic composite structures. The inverse multi-quadric (MQ), Gaussian, and MQ RBFs are used to test the stability of the RBF collocation method in periodic structures. Much useful information is obtained. Due to the merits of the RBF collocation method, the general form in this paper can easily be applied in the high dimensional problems analysis. The stability is fully discussed with different RBFs. The choice of the shape parameter and the effects of the knot number are presented.     

人工神经网络在溶解度预测方面的应用

溶解度作为一项重要的物化指标,一直是化学学科的研究重点。然而,通过实验测量获得数据耗时费力,因此,科研人员建立了多种理论方法来进行估算,其中,人工神经网络因其能够关联复杂的多变量情况而受到广泛关注。本文综述了人工神经网络在物质溶解度预测方面的应用,介绍了应用最广泛的3种神经网络(BP神经网络、小波神经网络、径向基神经网络)的模型结构、预测方法和预测优势,探讨了神经网络的不足以及改进方法。文章最后对神经网络在物质溶解度预测方面的发展前景进行了展望。与其他方法相比,人工神经网络技术在物质溶解度预测方面具有预测结果精确度高、操作简单等特点,具有广阔的应用前景,但输入变量选择、隐含层节点数确定、避免局部最优等问题还需逐步建立系统的理论指导。     

A novel CO2 gas analyzer based on IR absorption

Carbon dioxide (CO₂) gas analyzer can be widely used in many fields. A novel CO₂ gas analyzer based on infrared ray (IR) absorption is presented sufficiently in this paper. Applying Lambert–Beer Law, a novel space-double-beam optical structure is established successfully. The optical structure includes an IR source, a gas cell, a bandpass filter with a transmission wavelength at 4.26 μm, another bandpass filter with a transmission wavelength at 3.9 μm, and two IR detectors. Based on Redial Basic Function (RBF) artificial neural network, the measuring model of IR CO₂ analyzer is established with a high accuracy. A dynamic compensation filter is effectively designed to improve the dynamic characteristic of the IR CO₂ analyzer without gas pump. The IR CO₂ analyzer possesses the advantages of high accuracy and mechanical reliability with small volume, lightweight, and low-power consumption. Therefore, it can be used in such relevant fields as environmental protection, processing control, chemical analysis, medical diagnosis, and space environmental and control systems.     

Integrated radial basis functions‐based differential quadrature method and its performance

In this paper, indirect radial basis function networks (IRBFN) proposed by Nam and Tranh (Neural Networks 2001; 14 (2):185–199; Appl. Math. Modelling 2003; 27 :197–220) are incorporated into the differential quadrature (DQ) approximation of derivatives. For simplicity, this new variant of RBF‐DQ approach is named as iRBF‐DQ method. The proposed approach is validated by its application to solve the one‐dimensional Burger's equation, and simulate natural convection in a concentric annulus by solving Navier–Stokes equations. It was found that as compared to the benchmark data, the iRBF‐DQ approach can provide more accurate results than the original RBF‐DQ method. Copyright © 2006 John Wiley & Sons, Ltd.     

一种利用RBF神经网络的传感器建模新方法

介绍了基于RBF神经网络的传感器建模新方法，其网络结构简单、收敛性好、计算机仿真结果显示该模型能同时传感器的温度补偿和非线性校正。     

基于一种新型聚类算法的RBF神经网络混沌时间序列预测

运用两阶段学习方法构建径向基函数(RBF)神经网络模型预测混沌时间序列.在利用非监督学习算法确定网络隐层中心时,提出了一种基于高斯基的距离度量,并联合输入输出聚类的策略.基于Fisher可分离率设计高斯基距离度量中的惩罚因子,可以提高聚类的性能.而输入输出聚类策略的引入,建立了聚类性能与网络预测性能之间的联系.因此,根据本文方法构建的网络模型,一方面可以加快网络训练的速度,另一方面可以提高预测性能.将该方法对Mackey-Glass, Lorenz和Logistic混沌时间序列进行了预测仿真研究,仿真结果表明了该方法的有效性. 关键词： 混沌时间序列 预测 径向基神经网络 聚类