基于加权模糊支持向量描述的旋转机械故障分类

基于支持向量数据描述良好的分类性能,针对旋转机械故障诊断中故障样本获取的特点,提出了基于正负类样本的加权模糊支持向量数据描述多类分类器,不仅考虑了正类样本,而且也充分考虑了负类样本对分类结果的影响.利用模拟故障样本对系统进行了实验,结果表明提出的方法在系统中具有良好的分类能力.     

基于改进FOA算法优化WKELM的医疗滚动轴承故障诊断技术研究

针对传统小波核极限学习机（Extreme Learning Machine-ELM）应用于医疗滚动轴承故障诊断中识别精度不高且训练速度慢的一系列问题的出现,并针对性的想出一种更好的对滚动转轴发生的故障进行识别的办法,通过对小波核极限学习机算法进行改进的方法。该方法运用改进果蝇算法（LGMS-Fruit-flying Optimization Algorithm, LGMS-FOA）优化小波核极限学习机中的正则化系数和小波核函数中的参数。采用的方法是变分模态分解（Variational Mode Decomposition-VMD）,通过这种方法能够对滚动轴承的故障信号分解为含有故障信息的各模态分量从而提取到故障特征。通过与其他三种算法的实验结果对比证明,基于LGMS-FOA-WKELM的滚动轴承故障诊断方法的识别精度更高且训练时间更短。     

基于SVDD的轻轨锚固螺杆故障诊断研究

为了实现对重庆市轻轨轨道梁锚固螺杆的故障检测,提出了一种基于支持向量数据描述的锚固螺杆故障诊断方法,该方法只需要正常螺杆样本,且不需要对原始数据进行特征提取,就可以建立单值分类器,解决了缺少故障螺杆样本的难题。通过与常见的三种单值分类方法比较,表明SVDD分类器具有很好的分类效果和计算效率,能较好地区分正常螺杆和非正常螺杆,为轻轨锚固螺杆故障检测提供了新的诊断方法。     

基于主元分析的支持向量数据描述机械故障诊断

针对机械故障诊断缺乏故障样本的问题,提出了故障诊断单值分类法--支持向量数据描述法(SVDD).这种方法只需要正常运行状态的数据样本,就可以建立单值分类器,区分出正常和异常状态.试验以轴承为研究对象,采用主元分析法(PCA)作数据前处理,提取振动信号的统计特征值,得到的主元特征指标输入到SVDD分类器进行训练和测试.试验结果表明,PCA对正常和故障样本有较大的区分度,SVDD分类器能很好的分辨出轴承正常和故障状态,并且对未知故障有良好的识别能力.     

协同训练算法在滚动轴承故障诊断中的应用

针对单个分类器方法在滚动轴承故障诊断中精度较低、故障样本标记稀缺、特征空间维度高等问题，提出一种将协同训练与集成学习相结合的Co-Forest轴承故障诊断算法。Co-Forest是半监督学习中的协同训练算法，包含多个基分类器，通过投票实现协同训练中的置信度估算。从滚动轴承的振动信号中提取时域、频域特征指标。利用少量带标签和大量未标记样本重复地训练基分类器。集成基分类器，实现对滚动轴承故障的诊断。实验结果表明，与同类型的协同训练算法（Co-Training、Tri-Training）相比，Co-Forest算法在轴承故障诊断中具有更高的正确率，与当前针对特征向量高维、标记样本稀缺问题的ISS-LPP算法，SS-LLTSA算法相比，Co-Forest算法在保持很高诊断正确率的情况下，不需要降维、参数设置简单，具有一定的实际应用价值。     

基于MW-KECA与变量贡献SVDD的间歇过程故障检测系统

针对间歇过程的非线性和时变性特点以及故障易误报的问题,提出了一种将移动窗-核熵成分分析(MW-KECA)故障监测与基于变量贡献的支持向量数据描述(SVDD)故障诊断集合而成的故障检测系统。MW-KECA方法构建局部模型能有效处理数据的时变性,同时保留KECA优秀的非线性处理能力。故障诊断中以各变量对CS统计量-向量间角度关系指标的贡献作为输入数据来构建SVDD分类器,相较于原始数据,故障贡献能够突出同类相似信息和异类差异信息。通过青霉素发酵仿真实验,验证了检测系统在监测准确性与故障识别率上都有良好效果,证明了该检测系统的有效性。     

Sparse random projection-based hyperdisk classifier for bevel gearbox fault diagnosis

The fault diagnosis of bevel gearbox is of great significance. At present, the commonly used methods are based on pattern recognition, such as support vector machine, convex hull classifier and hyperdisk classifier. However, the number of elements in the kernel matrix of these kernel function-based classification methods increases squarely with the data size, resulting in intolerable training time. Based on this, a sparse random projection-based hyperdisk classifier model is proposed. The proposed method has the following novelties: First, based on sparse random projection and the geometrical characteristics of the hyperdisk model, a method is designed to efficiently screen out the core samples, and these samples are given different weights in this process. Second, the proposed method introduces slack variables and the dynamic penalty parameter to obtain a hyperdisk model with more reasonable boundary. Last, a strategy is developed to minimize the adverse effects of imbalanced training data. The effectiveness and applicability of the proposed method are verified on bevel gearbox fault data. The experimental results show that compared with other classifiers, the proposed method can greatly reduce the training time while guaranteeing a high classification accuracy. What’s more, it has better performance and efficiency in fault diagnosis with imbalanced training data.     

基于可拒识-双层支持向量分类器的微钙化点检测

为克服医学图像微钙化点检测中假阳性高的缺点,构造了一种带拒识能力的双层支持向量模型分类器,用于钙化点检测.检测时,首先利用基于最大间隔超平面的支持向量分类器（SVC）对输入模式进行分类判决;然后通过求取真实钙化点样本特征空间最小的包含球形边界来得到钙化点样本的球形支持向量域表示（SVDD）;接着利用钙化点的支持向量域表示对输入模式进行拒识或接受处理;最后利用SVC与SVDD两个分类器的结果来进行综合判决.仿真实验结果表明,该算法在不影响微钙化点的检出率的情况下,可部分解决假阳性高的问题.     

一种基于Morlet小波核的约简支持向量机

针对支持向量机（SVM）的训练数据量仅局限于较小样本集的问题,结合Morlet小波核函数,提出了一种基于Morlet小波核的约倚支持向量机（MWRSVM—DC）．算法的核心是通过密度聚类寻找聚类中每个簇的边缘点作为约倚集合,并利用该约倚集合寻找支持向量．实验表明,利用小波核,该算法不仅提高了分类的准确率,而且提高了整体分类效率．     

Intelligent fault diagnosis for rolling bearings based on graph shift regularization with directed graphs

Graph shift regularization is a new and effective graph-based semi-supervised classification method, but its performance is closely related to the representation graphs. Since directed graphs can convey more information about the relationship between vertices than undirected graphs, an intelligent method called graph shift regularization with directed graphs (GSR-D) is presented for fault diagnosis of rolling bearings. For greatly improving the diagnosis performance of GSR-D, a directed and weighted k-nearest neighbor graph is first constructed by treating each sample (i.e., each vibration signal segment) as a vertex, in which the similarity between samples is measured by cosine distance instead of the commonly used Euclidean distance, and the edge weights are also defined by cosine distance instead of the commonly used heat kernel. Then, the labels of samples are considered as the graph signals indexed by the vertices of the representation graph. Finally, the states of unlabeled samples are predicted by finding a graph signal that has minimal total variation and satisfies the constraint given by labeled samples as much as possible. Experimental results indicate that GSR-D is better and more stable than the standard convolutional neural network and support vector machine in rolling bearing fault diagnosis, and GSR-D only has two tuning parameters with certain robustness.     

面向新异检测的启发式约减支持向量数据描述

针对支持向量数据描述(SVDD)单类分类方法运算复杂度高的缺点,提出一种启发式约减支持向量数据描述(HR-SVDD)方法。以启发的方式从原有训练集中筛选出部分样本构成约减训练集,对约减训练集进行二次规划解算,得到支持向量和决策边界。通过不同宽度系数高斯核SVDD特征的讨论,证明了HR-SVDD的有效性。人工数据集和真实数据集上的实验结果表明, HR-SVDD分类精度与传统支持向量数据描述相当,但具有更快的运算速度和更小的内存占用。     

基于改进的PCM支持向量描述多类分类器

基于支持向量数据描述和改进的可能性c-均值聚类算法,提出了一种模糊的多类分类学习机.首先通过一个改进的PCM算法来计算每个样本对于每类的权值矩阵,该权值也反映了该样本对某类的重要程度;然后将该权值矩阵应用到支持向量数据描述方法中,并对样本进行训练;最后给出了一个针对多类分类的分类规则(函数),并从理论上证明该分类规则满足贝叶斯优化决策理论.通过对比实验分析,本文提出的算法在分类精度和训练时间上都有较大的改善.     

KSVDD及其在拒识判别中的应用

为提高支持向量域分类器(SVDC)的分类精度和鲁棒性,提出基于K近邻(KNN)和支持向量域描述(SVDD)的分类器KNN-SVDD (KSVDD)。该分类器对单类内部的样本采用SVDD的判别准则,对类交叉区域及描述边界外的样本采用KNN的判别准则。通过拒绝描述边界外的样本,KSVDD可应用于拒识判别。UCI数据集上的数值实验表明,KSVDD分类精度与支持向量机(SVM)相当且均比SVDC高,训练时间比SVM短,鲁棒性强,在拒识判别中有良好表现。     

基于概率的支持向量数据描述方法

针对目前概率机器学习方法在解决概率问题时具有较高的复杂度,而传统的支持向量数据描述（SVDD）作为一种核密度估计方法只能判断测试样本是否属于该类等问题,提出一种基于概率的支持向量数据描述方法。首先,利用传统的SVDD方法分别得到两类数据的数据描述,计算测试样本到超球体的距离;然后,构造一个将距离转换为概率的函数,提出一种基于概率的SVDD方法;同时,使用Bagging算法进行集成,进一步提高数据描述的性能。借鉴分类场景,将所提方法与传统的SVDD方法在Gunnar Raetsch的13种基准数据集上进行实验,实验结果表明,所提方法在准确率和F1值上优于传统的SVDD方法,并且其数据描述的性能有所提升。     

Vibrant fault diagnosis for hydroelectric generator units with a new combination of rough sets and support vector machine

The fault diagnosis for hydroelectric generator unit (HGU) is significant to prevent dangerous accidents from occurring and to improve economic efficiency. The faults of HGU involve overlapping fault patterns which may denote a kind of faults in the early stage or a subset of samples that caused by multi-fault. But until now it has not been considered in the traditional classifier of fault diagnosis for HGU. In this paper, a novel classifier combined rough sets and support vector machine is proposed and applied in the fault diagnosis for HGU. Instead of classifying the patterns directly, the fault patterns lying in the overlapped region are extracted firstly. Then, upper and lower approximations of each class are defined on the basis of rough set technique. Next, for the fault patterns lying in the overlapped region, the reliability they belong to a certain class is calculated. At last, the proposed method is successfully applied in analyzing an international standard data set, as well as diagnosing the vibrant faults of a HGU. The results show that the proposed classifier can more properly describe the complex map between the faults and their symptoms, and is suitable to fault diagnosis for HGU.     

基于DAGSVM的高炉故障诊断研究

针对高炉故障诊断智能化程度低，对操作人员技术水平要求高等不足，提出了基于支持向量机的多类分类故障诊断方法．根据统计学原理，使用核函数将样本映射到高维空间进行训练．综合各种核函数的测试准确率，得到解决该问题的最佳核函数．通过比较不同的多类分类算法，提出了基于DAGSVM的诊断模型．实验结果表明该算法具有较高的识别准确率．