基于核相似性差异最大化的支持向量机参数选择算法

针对支持向量机(SVM)参数选择问题,通过分析SVM近似网络模型及分类原理,提出一种基于核相似性差异最大化的高斯核参数快速选择算法(MSD)。同时,将MSD算法与基于交叉验证的参数搜索算法相结合,构成一种复合SVM参数选择算法(MSD-GS),实现核参数与正则化参数的快速优选。UCI数据的仿真实验表明该算法具有参数选择准确、简便快速、无需数据先验知识等优点,参数选择效果甚至优于遍历式指数网格搜索算法。优选出的参数组合能够使SVM具有较高的泛化性能。     

Parameter selection of support vector regression based on hybrid optimization algorithm and its application

Choosing optimal parameters for support vector regression （SVR） is an important step in SVR. design, which strongly affects the pefformance of SVR. In this paper, based on the analysis of influence of SVR parameters on generalization error, a new approach with two steps is proposed for selecting SVR parameters, First the kernel function and SVM parameters are optimized roughly through genetic algorithm, then the kernel parameter is finely adjusted by local linear search, This approach has been successfully applied to the prediction model of the sulfur content in hot metal. The experiment results show that the proposed approach can yield better generalization performance of SVR than other methods,     

Parameter selection of support vector regression based on hybrid optimization algorithm and its application

Choosing optimal parameters for support vector regression (SVR) is an important step in SVR design, which strongly affects the performance of SVR. In this paper, based on the analysis of influence of SVR parameters on generalization error, a new approach with two steps is proposed for selecting SVR parameters . First the kernel function and SVM parameters are optimized roughly through genetic algorithm, then the kernel parameter is finely adjusted by local linear search. This approach has been successfully applied to the prediction model of the sulfur content in hot metal. The experiment results show that the proposed approach can yield better generalization performance of SVR than other methods.     

通过全局核降低高斯核的局部风险与基于遗传算法的两阶段模型选择

支持向量分类时,由于样本分布的不均匀性,单宽度的高斯核会在空间的稠密区域产生过学习现象,在稀疏区域产生欠学习现象,即存在局部风险.针对于此,构造了一个全局性次核来降低高斯核产生的局部风险.形成的混合核称为主次核.利用幂级数构造性地给出并证明了主次核的正定性条件,进一步提出了基于遗传算法的两阶段模型选择算法来优化主次核的参数.实验验证了主次核和模型选择法的优越性.     

一种基于PSO的RBF-SVM模型优化新方法

针对使用径向基核函数的支持向量机,采用粒子群优化方法实现模型优化.基于训练集中样本之间的最近平均距离和最远平均距离,给出参数σ的取值空间,从而减小了超参数搜索的范围,并采用对数刻度进一步提高粒子群优化方法的参数搜索效率.与遗传算法和网格法的对比实验表明,所提出的方法收敛速度更快,得出的超参数更优.     

基于核路径算法的支持向量回归机参数选择

参数选择是支持向量机研究领域的重要问题.针对核参数的选择,提出一种基于二分法的核参数解路径算法.由于解为核参数的非线性光滑函数,该算法随着参数的更新,可以在已有参数得出的解的基础上通过更新公式进行推导计算,从而求得当前参数所对应的解,其目标函数的极值所对应的参数值即为最优参数解.该算法可以快速地求得最优参数.将该方法应用于双酚A生产过程的质量指标软测量建模,仿真结果表明了该算法的可行性和有效性.     

基于改进秃鹰搜索算法的同步优化特征选择

针对传统支持向量机在封装式特征选择中分类效果差、子集选取冗余、计算性能易受核函数参数影响的不足,利用元启发式优化算法对其进行同步优化.首先利用莱维飞行策略和模拟退火机制对秃鹰搜索算法的局部搜索能力与勘探利用解空间能力进行改进,通过标准函数的测试结果验证其改进的有效性;其次将支持向量机核函数参数作为待优化目标,利用改进后...     

Learning general Gaussian kernel hyperparameters of SVMs using optimization on symmetric positive-definite matrices manifold

We propose a new method for general Gaussian kernel hyperparameter optimization for support vector machines classification. The hyperparameters are constrained to lie on a differentiable manifold. The proposed optimization technique is based on a gradient-like descent algorithm adapted to the geometrical structure of the manifold of symmetric positive-definite matrices. We compare the performance of our approach with the classical support vector machine for classification and with other methods of the state of the art on toy data and on real world data sets.     

Feature selection and parameter optimization for support vector machines: A new approach based on genetic algorithm with feature chromosomes

Support vector machines (SVM) are an emerging data classification technique with many diverse applications. The feature subset selection, along with the parameter setting in the SVM training procedure significantly influences the classification accuracy. In this paper, the asymptotic behaviors of support vector machines are fused with genetic algorithm (GA) and the feature chromosomes are generated, which thereby directs the search of genetic algorithm to the straight line of optimal generalization error in the superparameter space. On this basis, a new approach based on genetic algorithm with feature chromosomes, termed GA with feature chromosomes, is proposed to simultaneously optimize the feature subset and the parameters for SVM.To evaluate the proposed approach, the experiment adopts several real world datasets from the UCI database and from the Benchmark database. Compared with the GA without feature chromosomes, the grid search, and other approaches, the proposed approach not only has higher classification accuracy and smaller feature subsets, but also has fewer processing time.     

一种确定高斯核模型参数的新方法

支持向量机中核函数及其参数的选择非常重要，该文提出了一种利用支持向量之间的距离求取高斯核函数参数的有效方法。该方法充分利用了支持向量机方法的最优判别函数仅仅与支持向量有关，并且支持向量为高斯核中心的特点。实验结果表明，该方法较好地反映了图像特征的本质，解决了高斯核函数参数在实际使用中不易确定的问题。     

基于量子遗传算法的XML聚类方法

主要用模式分析的核方法与量子遗传算法相结合研究XML聚类,提出了一种基于量子遗传算法混合核聚算法的XML文档聚类新方法。该方法先对XML文档约简,以频繁标签序列建立向量空间核的核矩阵,用高斯核函数求解初始聚类和聚类中心,然后用初始聚类中心构造量子遗传算法的初始种群,通过量子遗传算法与核聚算法相结合求得全局最优解的聚类。实验结果表明,使用该算法的聚类比改进的核聚算法、K均值算法等单一方法具有良好的收敛性、稳定性和更高的全局最优。     

Complex system fault diagnosis based on a fuzzy robust wavelet support vector classifier and an adaptive Gaussian particle swarm optimization

This paper proposes a robust loss function that penalizes hybrid noise (i.e., Gaussian noise, singularity points, and larger magnitude noise) in a complex fuzzy fault-diagnosis system. A mapping relationship between fuzzy numbers and crisp real numbers that allows a fuzzy sample set to be transformed into a crisp real sample set is also presented. Furthermore, the paper proposes a novel fuzzy robust wavelet support vector classifier (FRWSVC) based on a wavelet base function and develops an adaptive Gaussian particle swarm optimization (AGPSO) algorithm to seek the optimal unknown parameter of the FRWSVC. The results of experiments that apply the hybrid diagnosis model based on the FRWSVC and the AGPSO algorithm to fault diagnosis demonstrate that it is both feasible and effective. Tests comparing the method proposed in this paper against other fuzzy support vector classifier (FSVC) machines show that it outperforms them.     

基于支持向量机的NSCT域自适应图像水印算法

为进一步提升支持向量机水印算法鲁棒性,提出基于支持向量机的NSCT域自适应图像水印算法。主要思想是根据图像自身特征生成自适应嵌入水印序列,利用模糊核聚类和支持向量机对NSCT低频系数进行分类,选取适合嵌入水印的低频系数,然后利用支持向量机建立NSCT邻域系数的关系模型,自适应完成水印嵌入。算法具有良好的不可感知性、安全性,并通过嵌入自适应水印达到全盲水印检测。实验结果表明,提出算法对高斯噪声、椒盐噪声、低通滤波、中值滤波、均值滤波、JPEG、旋转、平移和尺寸缩放有很强的鲁棒性。     

基于QGA-LSSVM的醋酸乙烯聚合率软测量建模

针对最小二乘支持向量机(LS-SVM)在建立醋酸乙烯(VAC)聚合率软测量模型过程中最优模型参数的选择问题,提出了利用一种量子遗传算法来自动选取LS-SVM模型正则化参数和核函数参数的方法;把LS-SVM模型参数的选择问题转化为优化问题,利用全局搜索能力强的量子遗传算法优化LS-SVM建模过程的重要参数,建立了基于QGA-LSSVM方法的VAC聚合率软测量模型;仿真结果表明:与已有的神经网络和支持向量机软测量方法相比,该模型泛化能力强,精度高,更有利于醋酸乙烯聚合率测量工程实际运用。     

基于遗传优化最小二乘支持向量机的导引头故障诊断

为提高导引头故障诊断准确率,提出了一种采用改进遗传算法优化的最小二乘支持向量机（LSSVM）构造导引头多故障分类模型的方法。该方法基于一对一策略及改进的投票法建立两层LSSVM多故障分类器,并利用一种自适应变步长搜索策略改进的遗传算法对LSSVM的核参数和正则化参数进行自动优选。通过对某型导引头实测数据的仿真并和标准SVM及BP神经网络诊断方法相比较,结果表明该方法具有更高诊断准确率和计算效率。     

基于AIC的RVM核参数选择方法及其应用

在采用高斯径向基函数的相关向量机(RVM)回归模型中,核参数与模型性能之间关系复杂,针对如何确定RVM核参数的问题,提出一种基于AIC准则选择RVM的核参数的方法。首先基于Akaike Information Criterion (AIC)思想,得出一种新的统计量Q,同时将Q作为适应度函数;然后利用微分进化算法（Differential Evolution Algorithm,DE）对核参数进行寻优,以此选择确定核参数;最后利用该算法建立RVM回归模型对黄金价格进行短期预测。实验结果表明,该模型较传统方法建立的预测模型具有更高的拟合精度和更好的泛化能力,进一步证明基于AIC准则选择RVM的核参数的方法的可行性和有效性。     

基于遗传算法对支持向量机模型中参数优化

支持向量机是基于统计学习理论的结构风险最小化原理基础上提出来的一种学习算法,其在理论上保证了模型的最大泛化能力.针对支持向量机结构参数的选取在没有理论支持,选取又比较困难的情况下,对影响模型分类能力的相关参数进行了研究,提出了一种基于遗传算法和十折交叉检验相结合的遗传支持向量机(GA-SVM)算法,利用遗传算法的全局搜索特性得到支持向量机(SVM)的最优参数值,并用算例表明了此算法有效提高了分类的精度和效率.     

A new intelligent diagnosis system for the heart valve diseases by using genetic-SVM classifier

In this study, an intelligent system based on genetic-support vector machines (GSVM) approach is presented for classification of the Doppler signals of the heart valve diseases. This intelligent system deals with combination of the feature extraction and classification from measured Doppler signal waveforms at the heart valve using the Doppler ultrasound. GSVM is used in this study for diagnosis of the heart valve diseases. The GSVM selects of most appropriate wavelet filter type for problem, wavelet entropy parameter, the optimal kernel function type, kernel function parameter, and soft margin constant C penalty parameter of support vector machines (SVM) classifier. The performance of the GSVM system proposed in this study is evaluated in 215 samples. The test results show that this GSVM system is effective to detect Doppler heart sounds. The averaged rate of correct classification rate was about 95%.     

基于支持向量机的分段线性学习方法

1.引言包括感知器、神经网络等在内的学习方法都是基于经验风险最小(ERM)原则的,而在实际的基于小样本的学习系统中,这些学习方法在经验风险最小的情况下并不能保证期望风险最小化。对于线性不可分情况不能给出是否分段线性可分的可靠信息。如果简单地引入非线性变换,则容易导致过学习现象。这显然不是我们所希望的。     

基于高斯核的SVM的参数选择

基于高斯核的支持向量机应用很广泛,高斯核参数σ的选择对分类器性能影响很大,本文提出了从核函数性质和几何距离角度来选择参数σ,并且利用高斯函数的麦克劳林展开解决了参数σ的优化选择问题。实验结果表明,该方法能较快地确定核函数参数σ,且 SVM 分类效果较好,解决了高斯核参数σ在实际应用中不易确定的问题。