极限学习机的分类器集成模型研究

将极限学习机算法与旋转森林算法相结合,提出了以ELM算法为基分类器并以旋转森林算法为框架的RF-ELM集成学习模型。在8个数据集上进行了3组预测实验,根据实验结果讨论了ELM算法中隐含层神经元个数对预测结果的影响以及单个ELM模型预测结果不稳定的缺陷;将RF-ELM模型与单ELM模型和基于Bagging算法集成的ELM模型相比较,由稳定性和预测精度的两组对比实验的实验结果表明,对ELM的集成学习可以有效地提高ELM模型的性能,且RF-ELM模型较其他两个模型具有更好的稳定性和更高的准确率,验证了RF-ELM是一种有效的ELM集成学习模型。     

基于风格感知的无监督领域适应算法

近年来,神经机器翻译的译文质量取得了显著的进步,但是其在训练过程中严重依赖平行的双语句子对.然而对于电子商务领域来说,平行资源是稀缺的,此外,文化的不同导致产品信息表达存在风格差异.为了解决这两个问题,提出了一种基于风格感知的无监督领域适应算法,该算法在互训练方法中充分利用电子商务单语数据,同时引入拟知识蒸馏的方法处理...     

基于半监督极限学习机的轴承故障诊断

鉴于在实际的应用中滚动轴承的故障信号所属的类别往往是未知的,而且为了得到一定的测试数据需要花费大量的时间,甚至对机械设备造成了一些损害.利用极限学习机训练速度快且泛化能力强的特点,提出了一种基于半监督极限学习机的滚动轴承故障诊断方法,该方法允许在有少量带标签的轴承故障数据的情况下,将带标签的历史数据与新采集到的部分未带标签的数据一起用来训练得到一个最优的诊断模型.首先通过相空间重构将原始一维信号映射到一个高维的相空间,在相空间中提取初始的轴承特征集,然后将特征集输入半监督的极限学习机中进行训练和测试.实验结果表明,这种基于半监督算法的诊断模型简单,在神经元个数较少的情况下仍然具有很好的泛化能力,具有一定的应用价值.     

基于多输出极限学习机的快速一致性分类器

一致性分类器是建立在一致性预测基础上的分类器,其输出结果具有很高的可靠性,但由于计算框架的限制,学习的时间往往较长.为了加快学习速度,首次将一致性预测与多输出极限学习机相结合,提出基于两者的快速一致性分类算法.该算法利用了极限学习机,能够快速计算样本标签的留一交叉估计的特性,极大地加快了学习速度.算法复杂度分析表明,所提算法的计算复杂度与多输出极限学习机的算法复杂度相同,该算法继承了一致性预测的可靠性特征,即预测的错误率能够被显著性水平参数所控制.在10个公共数据集上的对比实验表明,所提算法具有极快的计算速度,且与其他常用一致性分类器相比,该算法的平均预测标签个数在某些数据集上更少,预测结果更有效.     

流形极限学习机自编码特征表示

极限学习机（ELM）作为一种无监督分类方法,具有学习速度快、泛化性能高、逼近能力好的优点。随着无监督学习的发展,将ELM与自动编码器集成已成为无标签数据集提取特征的新视角,如极限学习机自动编码器（ELM-AE）是一种无监督的神经网络,无需迭代即可找到代表原始样本和其学习过程的主要成分。其重建输入信号获取原始样本的主要特征,且考虑了原始数据的全局信息以避免信息的丢失,然而这类方法未考虑数据的固有流形结构即样本间的近邻结构关系。借鉴极限学习机自动编码器的思想,提出了一种基于流形的极限学习机自动编码器算法（M-ELM）。该算法是一种非线性无监督特征提取方法,结合流形学习保持数据的局部信息,且在特征提取过程中同时对相似度矩阵进行学习。通过在IRIS数据集、脑电数据集和基因表达数据集上进行实验,将该算法与其他无监督学习方法PCA、LPP、NPE、LE和ELM-AE算法经过[k]-means聚类后的准确率进行了比较,以表明该算法的有效性。     

选择性集成极限学习机分类器建模研究

极限学习机ELM(Extreme Learning Machine)具有训练过程极为快速的优点,但在实际分类应用中ELM分类器的分类精度和稳定性有时并不能满足要求。针对这一问题,在ELM用于分类时引入一种训练结果信息量评价指标来改进输出权值矩阵的求解方法,并增加隐层输出矩阵竞争机制来提高ELM的稳定性。为了进一步提高ELM的分类正确率,借鉴神经网络集成的理论,提出一种选择性集成ELM分类器。在集成方法中采用改进Bagging法并提出一种基于网络参数向量的相似度评价方法和选择性集成策略。最后通过UCI数据测试表明,同Bagging法和传统的全集成法相比,该方法拥有更为优秀的分类性能。     

基于极限学习机参数迁移的域适应算法

针对含少量标签样本的迁移学习问题,本文提出了基于极限学习机（Extreme learning machine,ELM）参数迁移的域适应算法,其核心思想是将目标域的ELM分类器参数投影到源域参数空间中,使其最大限度地与源域的分类器参数分布相同.此外,考虑到迁移中有可能带来负迁移的情况,在目标函数中引入正则项约束.本文算法与以往的域适应算法相比优势在于,其分类器参数以及转移矩阵是同时优化得到的,并且其目标函数求解过程相对简单.实验结果表明,与主流的域适应算法相比,本文算法在精度与效率上都表现出明显的优势.     

无逆矩阵在线序列极限学习机

无逆矩阵极限学习机只能以批量学习方式进行训练,将其拓展为无逆矩阵在线学习版本,提出了无逆矩阵在线序列极限学习机算法(IOS-ELM)。所提算法增加训练样本时,利用Sherman Morrison Woodbury公式对新增样本数据后的模型进行更新,直接计算出新增隐含层输出权重,避免对已经分析过的训练样本的输出权重进行重复计算。给出了所提IOS-ELM算法的详细推导过程。在不同类型和大小的数据集上的实验结果表明,所提IOS-ELM算法非常适合在线方式逐步生成的数据集,在快速学习和性能方面都有很好的表现。     

基于半监督极限学习机的增量式定位算法

针对目前室内指纹定位算法存在实时性差、对动态环境适应性不足的问题,提出一种新的基于半监督极限学习机的定位算法.该算法首先通过半监督极限学习机建立初始化位置估计模型,然后利用新增的半标记数据对原定位模型进行动态调整,最后为新增训练数据分配合适惩罚权重,使模型具有时效机制.仿真结果表明,该定位算法在保证定位实时性的同时提高了对动态环境的适应性.     

基于健壮主成分分类器的无监督异常检测方法研究

入侵检测系统在训练过程中需要大量有标识的监督数据进行学习,不利于其应用和推广,经典主成分分析方法对离群数据非常敏感,进而导致分类准确性的下降。为了解决该问题,提出了一种基于健壮主成分分类器的方法,得到被离群数据干扰较少的主成分。根据主成分空间距离和数据重构误差构建异常检测模型。实验表明：该方法能够有效检测未知入侵,在检测率、误警率方面都达到较满意的结果。     

RKHS subspace domain adaption via minimum distribution gap

Pattern Analysis and Applications - Subspace learning of Reproducing Kernel Hilbert Space (RKHS) is most popular among domain adaption applications. The key goal is to embed the source and target...     

Multi-context unsupervised domain adaption for HEp-2 cell classification using maximum partial classifier discrepancy

Identifying human epithelial-2 (HEp-2) cells in indirect immune fluorescence (IIF) is the most commonly used method for the diagnosis of autoimmune diseases. Although supervised deep learning networks have made remarkable progress on HEp-2 cell staining pattern classification, the high-performance relies on a large amount of labeled training data. Unfortunately, large-scale labeled datasets are scarce due to the expensive costs of labeling medical images. Therefore, we propose an unsupervised domain adaption (UDA) model, namely MDA-MPCD, to classify unlabeled HEp-2 cell samples. The proposed model involves two major aspects: (a) multi-context domain adaption (MDA) generator and (b) maximum partial classifier discrepancy (MPCD) architecture. The MDA generator can extract multi-context features from complex cell images while providing more comprehensive and diverse information for the classifier. The MPCD architecture, utilizing the mapping variation of feature transfer, focuses on the discrepancy from the cross-domain gap by separating the activations in the classifier. The proposed model dominates all comparison methods during evaluation, achieving 85.35% and 96.08% mean accuracy on two UDA tasks, respectively. Furthermore, the model is demonstrated to adapt from rich label domain to unlabeled domain by detailed ablation experiments and visualization results. The proposed MDA-MPCD has potential as a clinical scheme, enabling effective and accurate classification of HEp-2 cell staining pattern without labeling the target data.

     

Unsupervised visual domain adaptation via discriminative dictionary evolution

Pattern Analysis and Applications - This work focuses on unsupervised visual domain adaptation which is still challenging in visual recognition. Most of the attention has been dedicated to seeking...     

Unsupervised classification using polarimetric interferometric decomposition and Wishart classifier

ABSTRACT

In this paper, a decomposition scheme of the coherency matrix is presented to parse the information of polarimetric interferometric synthetic aperture radar (PolInSAR) images in detail. First, the decomposition method is improved by the polarimetric interferometric similarity parameter (PISP) to relief the overestimation occurred in the traditional four-component decomposition method. Second, after using the improved four-component decomposition results as the original inputs, the decomposition method is applied to retrieve scattering mechanisms or identify scatters, with the image separated into seven subsections. Finally, based on the modified decomposition results, the basic classification results are regarded as the feature training sets, and the Wishart classifier is then used as an optimized classification process. The applications of the decomposition and classification scheme are shown with typical representative L-band E-SAR images, which are used to show the robustness of the method, as well as with the first published airborne C-band PolInSAR data collected by the Institute of Electronics, Chinese Academy of Sciences, in November 2017. Experimental results demonstrate that the obtained decomposition and classification results are in good agreement with the actual physical scattering mechanisms.     

基于RSDE 的领域自适应概率密度估计方法

同一应用领域不同时间、地点或设备,采集的样本数据可能存在扰动、噪音或缺失,如何对样本数据集进行有效的预处理是其进一步应用的前提.针对上述问题,提出一种新的基于压缩集密度估计(RSDE)算法的领域自适应概率密度估计方法 A-RSDE,通过学习源域(训练域)知识,使目标域(测试域)概率密度估计更接近真实概率密度分布,并用基于近似最小包含球的核心集快速算法求解 A-RSDE,将其应用于大数据集密度估计. Benchmark 和 UCI 数据集上的实验表明,该算法具有较好的性能.     

聚类分片双支持向量域分类器

针对支持向量域分类器对大规模样本集的训练时间长且占用内存大的问题,构造聚类分片双支持向量域分类器。以均值聚类剖分原始空间,并选取密度指标大的样本作为初始聚类中心;对子空间构造双支持向量域分类器,根据样本与正负类最小包围超球的距离构造分段决策函数;定义样本的变尺度距离,以链接规则组合子空间的分类结果。数值实验表明,所提出算法的分类精度高且受参数变化的影响不大,分类时间短且随子空间数的增加而降低。     

Unsupervised instance selection via conjectural hyperrectangles

Neural Computing and Applications - Machine learning algorithms spend a lot of time processing data because they are not fast enough to commit huge data sets. Instance selection algorithms...     

Tracking Persons-of-Interest via Unsupervised Representation Adaptation

International Journal of Computer Vision - Multi-face tracking in unconstrained videos is a challenging problem as faces of one person often can appear drastically different in multiple shots due...     

Diversity-aware classifier ensemble selection via f-score

The primary effect of using a reduced number of classifiers is a reduction in the computational requirements during learning and classification time. In addition to this obvious result, research shows that the fusion of all available classifiers is not a guarantee of best performance but good results on the average. The much researched issue of whether it is more convenient to fuse or to select has become even more of interest in recent years with the development of the Online Boosting theory, where a limited set of classifiers is continuously updated as new inputs are observed and classifications performed. The concept of online classification has recently received significant interest in the computer vision community. Classifiers can be trained on the visual features of a target, casting the tracking problem into a binary classification one: distinguishing the target from the background.Here we discuss how to optimize the performance of a classifier ensemble employed for target tracking in video sequences. In particular, we propose the F-score measure as a novel means to select the members of the ensemble in a dynamic fashion. For each frame, the ensemble is built as a subset of a larger pool of classifiers selecting its members according to their F-score. We observed an overall increase in classification accuracy and a general tendency in redundancy reduction among the members of an f-score optimized ensemble. We carried out our experiments both on benchmark binary datasets and standard video sequences.