共查询到20条相似文献,搜索用时 0 毫秒
1.
2.
介绍和比较标准支持向量机(SVM)和最小二乘支持向量机(LS-SVM)基本原理的基础上,探讨了一种利用LS-SVM进行传感器动态误差补偿的方法,并给出了相应的过程和算法。与标准SVM补偿方法比较,该方法的优点是明显的:用等式约束代替标准SVM算法中的不等式约束,将求解二次规划问题转化为直接求解线性矩阵方程,在相同样本条件下,使得补偿器构造速度提高1~2个数量级。通过对SVM和LS-SVM传感器动态补偿的仿真分析和实验结果对比表明,在噪声条件下,LS-SVM方法的补偿误差约为SVM的40%。因此,LS-SVM补偿方法学习速度快,抗噪声干扰能力强,更适合传感器动态补偿。 相似文献
3.
ZHUJia-yuan ZHANGXi-bin ZHANGHeng-xi RENBo 《国际设备工程与管理》2004,9(2):97-102
A multi-layer adaptive optimizing parameters algorithm is developed for improving least squares support vector machines (LS-SVM), and a military aircraft life-cycle-cost (LCC) intelligent estimation model is proposed based on the improved LS-SVM. The intelligent cost estimation process is divided into three steps in the model. In the first step, a cost-drive-factor needs to be selected, which is significant for cost estimation. In the second step, military aircraft training samples within costs and cost-drive-factor set are obtained by the LS-SVM. Then the model can be used for new type aircraft cost estimation. Chinese military aircraft costs are estimated in the paper. The results show that the estimuted costs by the new model are closer to the true costs than that of the traditionally used methods. 相似文献
4.
针对小样本步态数据引起的分类器泛化能力差的问题,提出了基于支持向量机的步态分类方法.采集了24名青年和24名老年受试者的步态数据,提取24个步态特征训练支持向量机,采用交叉验证方法评估分类器的泛化性能.结果表明,本文提出的方法能够有效地对小样本步态数据分类,并且具有良好的泛化性.不同的核函数对分类性能影响较小.与传统反向传播学习算法的神经网络分类器进行了比较,支持向量机分类性能明显优于传统反向传播学习算法的神经网络.支持向量机在步态分类中具有广泛的应用前景. 相似文献
5.
6.
7.
8.
9.
SVM多类分类算法及其在故障诊断中的应用 总被引:4,自引:0,他引:4
支持向量机理论最初是针对两类问题线性可分的模式识别提出来的.在故障诊断领域,多类故障诊断是经常出现的问题.介绍了几种基于支持向量机的多类分类算法的原理,在此基础上提出一种权重二叉树多类分类算法,考虑故障状态和正常状态、重要故障和次要故障、常见故障和不常见故障之间的权重不同,适当把最佳分类面往远离重要的或常见的类别一方偏离一定的距离.在小样本情况下对转子模拟试验台典型的多类故障进行诊断,结果表明它比较符合实际工程要求. 相似文献
11.
为解决SVM在积雨云检测中的难题,本文构造了一种模糊支持向量机(FSVM),首先根据训练样本的分布特性,定义了相邻样本距离类中心的距离变化率,然后通过计算距离变化率来剔除训练集中可能的噪声与野值样本,从而有效克服了传统基于紧密度的FSVM在计算最小超球半径时易受噪声与野值干扰的缺点,使得所计算的隶属度能更好地反映不同样本的差异。实验结果表明,对于FY2D卫星云图,采用从不同通道所提取的光谱特征,本文方法的积雨云检测准确率与传统SVM和基于紧密度的FSVM相比,分别平均提高2%和1%,且具有更强的适应性及噪声鲁棒性。 相似文献
12.
综合分析了影响汽柴油消费需求的关键因素,并针对其具有自相关性、复杂性、数据量大等特点,采用主成分分析法对样本数据进行降维处理,形成新的样本集。对支持向量机预测模型进行改进,在其基础之上引入时序动态因子,将上年的汽柴油需求历史数据作为时序反馈因子引入模型,从而形成新的动态反馈拟合模型,建立相应的需求预测模型。对1996~2012年的汽柴油需求预测进行实例研究,并将本文中所提方法的预测结果与灰色GM(1,1)模型、BP神经网络模型进行对比分析。结果表明本文中的主成分分析与改进支持向量机预测方法相对于GM(1,1)模型其预测误差均值分别降低了72.7%和74.86%,相对于BP神经网络其预测误差均值分别降低了81.3%和8166%,从而证明了此方法的有效性和优越性。 相似文献
13.
支持向量机及其在机械故障诊断中的应用 总被引:4,自引:6,他引:4
支持向量机(SVM)是一种基于统计学习理论的新型机器学习方法,对小样本决策具有较好的学习推广性。对近年来支持向量机的研究进展及其在故障诊断中的应用做了简要介绍,讨论了支持向量机的特点和存在的问题,展望了其在机械故障诊断的研究前景。 相似文献
14.
15.
16.
17.
Syed Muhammad Saqlain Shah Tahir Afzal Malik Robina khatoon Syed Saqlain Hassan Faiz Ali Shah 《计算机、材料和连续体(英文)》2019,61(2):535-553
Classification of human actions under video surveillance is gaining a lot of attention from computer vision researchers. In this paper, we have presented methodology to recognize human behavior in thin crowd which may be very helpful in surveillance. Research have mostly focused the problem of human detection in thin crowd, overall behavior of the crowd and actions of individuals in video sequences. Vision based Human behavior modeling is a complex task as it involves human detection, tracking, classifying normal and abnormal behavior. The proposed methodology takes input video and applies Gaussian based segmentation technique followed by post processing through presenting hole filling algorithm i.e., fill hole inside objects algorithm. Human detection is performed by presenting human detection algorithm and then geometrical features from human skeleton are extracted using feature extraction algorithm. The classification task is achieved using binary and multi class support vector machines. The proposed technique is validated through accuracy, precision, recall and F-measure metrics. 相似文献
18.
With a view to gaining an in-depth assessment of the response of particleboards (PBs) to different in-service loading conditions, samples of high-density homogeneous PBs of sugarcane bagasse and castor oil polyurethane resin were manufactured and subjected to low velocity impacts using an instrumented drop weight impact tower and four different energy levels, namely 5, 10, 20 and 30 J. The prediction of the damage modes was assessed using Comsol Multiphysics\(^\circledR .\) In particular, the random distribution of the fibres and their lengths were reproduced through a robust model. The experimentally obtained dent depths due to the impactor were compared with the ones numerically simulated showing good agreement. The post-impact damage was evaluated by a simultaneous system of image acquisitions coming from two different sensors. In particular, thermograms were recorded during the heating up and cooling down phases, while the specklegrams were gathered one at room temperature (as reference) and the remaining during the cooling down phase. On one hand, the specklegrams were processed via a new software package named Ncorr v.1.2, which is an open-source subset-based 2D digital image correlation (DIC) package that combines modern DIC algorithms proposed in the literature with additional enhancements. On the other hand, the thermographic results linked to a square pulse were compared with those coming from the laser line thermography technique that heats a line-region on the surface of the sample instead of a spot. Surprisingly, both the vibrothermography and the line scanning thermography methods coupled with a robotized system show substantial advantages in the defect detection around the impacted zone. 相似文献
19.
20.
本文提出了一种基于支持向量机的坦克识别算法。在对图像预处理之后,运用颜色和纹理信息进行分割,采用基于数学形态学的算法求得边缘像素,提取具有RST不变性的轮廓特征向量,输入支持向量机进行训练和识别。将支持向量机与传统的人工神经网络的算法进行了对比实验,实验表明基于支持向量机的坦克识别算法具有更好的性能。 相似文献