低秩稀疏图嵌入的半监督特征选择

目的 特征降维是机器学习领域的热点研究问题。现有的低秩稀疏保持投影方法忽略了原始数据空间和降维后的低维空间之间的信息损失,且现有的方法不能有效处理少量有标签数据和大量无标签数据的情况,针对这两个问题,提出基于低秩稀疏图嵌入的半监督特征选择方法（LRSE）。方法 LRSE方法包含两步：第1步是充分利用有标签数据和无标签数据分别学习其低秩稀疏表示,第2步是在目标函数中同时考虑数据降维前后的信息差异和降维过程中的结构信息保持,其中通过最小化信息损失函数使数据中有用的信息尽可能地保留下来,将包含数据全局结构和内部几何结构的低秩稀疏图嵌入在低维空间中使得原始数据空间中的结构信息保留下来,从而能选择出更有判别性的特征。结果 将本文方法在6个公共数据集上进行测试,对降维后的数据采用KNN分类验证本文方法的分类准确率,并与其他现有的降维算法进行实验对比,本文方法分类准确率均有所提高,在其中的5个数据集上本文方法都有最高的分类准确率,其分类准确率分别在Wine数据集上比次高算法鲁棒非监督特征选择算法（RUFS）高11.19%,在Breast数据集上比次高算法RUFS高0.57%,在Orlraws10P数据集上比次高算法多聚类特征选择算法（MCFS）高1%,在Coil20数据集上比次高算法MCFS高1.07%,在数据集Orl64上比次高算法MCFS高2.5%。结论 本文提出的基于低秩稀疏图嵌入的半监督特征选择算法使得降维后的数据能最大限度地保留原始数据包含的信息,且能有效处理少量有标签样本和大量无标签样本的情况。实验结果表明,本文方法比现有算法的分类效果更好,此外,由于本文方法基于所有的特征都在线性流形上的假设,所以本文方法只适用于线性流形上的数据。     

局部样条嵌入的正交半监督子空间学习算法

为了更加准确地对图像进行聚类与分类,提出一种基于局部样条嵌入的正交半监督子空间学习算法.通过学习一个正交投影矩阵,使得训练样本中的标注数据经过投影矩阵降维后类间离散度尽量大,类内离散度尽量小;采用局部样条回归将局部低维嵌入坐标映射成全局低维嵌入坐标,使得被投影数据保持原有流形结构,并有效地利用有标注训练样本和未标注训练样本得到优化的图像表达方式.图像聚类与分类实验的结果表明了文中算法的有效性.     

基于核的半监督的局部保留投影降维方法

为能有效捕捉数据的非线性特征,特提出一种新的非线性数据降维算法——核半监督局部保留投影(KSSLPP)。该方法利用标记样本的标记信息及所有训练样本的结构重新定义了类间相似度和类内相似度,然后将原始数据映射到高维核空间,在核空间中最大化类间分离度,最小化类内分离度。该方法在核空间保持了数据的局部结构和全局结构,以及数据的标签信息。在Olivetti人脸库和UCI数据库中的对比实验验证了该算法的有效性。     

全局判别与局部稀疏保持HSI半监督特征提取

针对高光谱图像存在“维数灾难”的问题，提出一种全局判别与局部稀疏保持的高光谱图像半监督特征提取算法（GLSSFE）。该算法通过LDA算法的散度矩阵保存有类标样本的全局类内判别信息和全局类间判别信息，结合利用半监督PCA算法对有类标和无类标样本进行主成分分析，保存样本的全局结构；利用稀疏表示优化模型自适应揭示样本数据间的非线性结构，将局部类间判别权值和局部类内判别权值嵌入半监督LPP算法保留样本数据的局部结构，从而最大化同类样本的相似性和异类样本的差异性。通过1-NN和SVM两个分类器分别对Indian Pines和Pavia University两个公共高光谱图像数据集进行分类，验证所提特征提取方法的有效性。实验结果表明，该GLSSFE算法最高总体分类精度分别达到89.10%和92.09%，优于现有的特征提取算法，能有效地挖掘高光谱图像的全局特征和局部特征，极大地提升高光谱图像的地物分类效果。     

一种半监督流形学习的人脸识别方法

针对传统线性降维方法忽略数据局部结构特性的问题,提出了一种基于半监督流形学习的方法。针对人脸识别采用图像欧式距离来选择各样本点的K近邻,由此得到修改后无监督判别投影中的邻接矩阵,在传统的无监督判别投影中,融入类标签信息获得几何最优投影。通过在人脸库上的大量比较实验,验证了该方法的准确性和有效性。     

基于半监督典型相关分析的多视图维数约简

为了有效地在半监督多视图情景下进行维数约简,提出了使用非负低秩图进行标签传播的半监督典型相关分析方法。非负低秩图捕获的全局线性近邻可以利用直接邻居和间接可达邻居的信息维持全局簇结构,同时,低秩的性质可以保持图的压缩表示。当无标签样本通过标签传播算法获得估计的标签信息后,在每个视图上构建软标签矩阵和概率类内散度矩阵。然后,通过最大化不同视图同类样本间相关性的同时最小化每个视图低维特征空间类内变化来提升特征鉴别能力。实验表明所提方法比已有相关方法能够取得更好的识别性能且更鲁棒。     

基于相似度矩阵学习和矩阵校正的无监督多视角特征选择

多视角特征选择通过融合多个视角的信息获取具有代表性的特征子集,来提高分类、聚类等学习任务的效率。然而,描述对象的特征繁杂多样且相互关联,单一地从原始特征中选择特征子空间可以简单地解决维度问题,但无法有效获取数据内部存在的结构信息和特征关联信息,且固定使用相似度矩阵和投影矩阵易损失视角间的相关性。针对以上问题,提出了基于相似度矩阵学习和矩阵校正的无监督多视角特征选择(SMLMA)算法。该算法首先构造所有视角的相似度矩阵,通过流形学习得到一致相似度矩阵以及投影矩阵,最大程度地发现和保留多视角数据的结构信息;其次采用矩阵校正的方法,最大化相似度矩阵和核矩阵之间的相关性,合理利用不同视角之间的关联性,减少特征子集的信息冗余;最后,采用Armijo搜索方法快速得到收敛结果。在4个实验数据集Caltech-7,NUS-WIDE-OBJ,Toy Animal和MSRC-v1上的实验结果表明,相比单视角特征选择和部分多视角特征选择方法,所提算法在聚类任务上的准确率平均提高了约7.54%。其较好地保留了数据的结构信息和多视角之间特征的相关性,捕获了更多高质量的特征。     

一种基于极大熵的快速无监督线性降维方法

现实世界中高维数据无处不在,然而在高维数据中往往存在大量的冗余和噪声信息,这导致很多传统聚类算法在对高维数据聚类时不能获得很好的性能.实践中发现高维数据的类簇结构往往嵌入在较低维的子空间中.因而,降维成为挖掘高维数据类簇结构的关键技术.在众多降维方法中,基于图的降维方法是研究的热点.然而,大部分基于图的降维算法存在以下两个问题:(1)需要计算或者学习邻接图,计算复杂度高;(2)降维的过程中没有考虑降维后的用途.针对这两个问题,提出一种基于极大熵的快速无监督降维算法MEDR. MEDR算法融合线性投影和极大熵聚类模型,通过一种有效的迭代优化算法寻找高维数据嵌入在低维子空间的潜在最优类簇结构. MEDR算法不需事先输入邻接图,具有样本个数的线性时间复杂度.在真实数据集上的实验结果表明,与传统的降维方法相比, MEDR算法能够找到更好地将高维数据投影到低维子空间的投影矩阵,使投影后的数据有利于聚类.     

基于图像矩阵判别局部保持投影的人脸识别

提出一种基于图像矩阵判别局部保持投影的人脸识别方法。图像矩阵判别局部保持投影是在局部保持投影基础上进行了扩展,考虑了类标签信息并在其目标函数中增加类间散度约束,使得求解的特征更具判别性。另外,图像矩阵判别局部保持投影是直接处理图像矩阵而不需要将矩阵转化为向量,保留了像素间的空间位置关系,避免了奇异性问题。实验结果表明该方法是有效的。     

基于二维邻域保持判别嵌入的人脸识别

提出二维邻域保持判别嵌入(2DNPDE)算法,该算法是一种有监督的基于二维图像矩阵的特征提取算法.为表示样本的类内邻域结构和类间距离关系,分别构建类内邻接矩阵和类间相似度矩阵.2DNPDE所获得的投影空间不但使不同类数据点的低维嵌入相互分离,而且保留同类样本的邻域结构和不同类样本的距离关系.在ORL和AR人脸数据库上的实验表明,该算法具有更好的识别效果.     

Multi-view clustering via simultaneous weighting on views and features

In big data era, more and more data are collected from multiple views, each of which reflect distinct perspectives of the data. Many multi-view data are accompanied by incompatible views and high dimension, both of which bring challenges for multi-view clustering. This paper proposes a strategy of simultaneous weighting on view and feature to discriminate their importance. Each feature of multi-view data is given bi-level weights to express its importance in feature level and view level, respectively. Furthermore, we implements the proposed weighting method in the classical k-means algorithm to conduct multi-view clustering task. An efficient gradient-based optimization algorithm is embedded into k-means algorithm to compute the bi-level weights automatically. Also, the convergence of the proposed weight updating method is proved by theoretical analysis. In experimental evaluation, synthetic datasets with varied noise and missing-value are created to investigate the robustness of the proposed approach. Then, the proposed approach is also compared with five state-of-the-art algorithms on three real-world datasets. The experiments show that the proposed method compares very favourably against the other methods.     

A novel ensemble construction method for multi-view data using random cross-view correlation between within-class examples

Correlated information between multiple views can provide useful information for building robust classifiers. One way to extract correlated features from different views is using canonical correlation analysis (CCA). However, CCA is an unsupervised method and can not preserve discriminant information in feature extraction. In this paper, we first incorporate discriminant information into CCA by using random cross-view correlations between within-class examples. Because of the random property, we can construct a lot of feature extractors based on CCA and random correlation. So furthermore, we fuse those feature extractors and propose a novel method called random correlation ensemble (RCE) for multi-view ensemble learning. We compare RCE with existing multi-view feature extraction methods including CCA and discriminant CCA (DCCA) which use all cross-view correlations between within-class examples, as well as the trivial ensembles of CCA and DCCA which adopt standard bagging and boosting strategies for ensemble learning. Experimental results on several multi-view data sets validate the effectiveness of the proposed method.     

Frobenius norm-regularized robust graph learning for multi-view subspace clustering

Graph learning methods have been widely used for multi-view clustering. However, such methods have the following challenges: (1) they usually perform simple fusion of fixed similarity graph matrices, ignoring its essential structure. (2) they are sensitive to noise and outliers because they usually learn the similarity matrix from the raw features. To solve these problems, we propose a novel multi-view subspace clustering method named Frobenius norm-regularized robust graph learning (RGL), which inherits desirable advantages (noise robustness and local information preservation) from the subspace clustering and manifold learning. Specifically, RGL uses Frobenius norm constraint and adjacency similarity learning to simultaneously explore the global information and local similarity of views. Furthermore, the l_2,1 norm is imposed on the error matrix to remove the disturbance of noise and outliers. An effectively iterative algorithm is designed to solve the RGL model by the alternation direction method of multipliers. Extensive experiments on nine benchmark databases show the clear advantage of the proposed method over fifteen state-of-the-art clustering methods.

     

Adaptive partial graph learning and fusion for incomplete multi-view clustering

Most of existing multi-view clustering methods assume that different feature views of data are fully observed. However, it is common that only portions of data features can be obtained in many practical applications. The presence of incomplete feature views hinders the performance of the conventional multi-view clustering methods to a large extent. Recently proposed incomplete multi-view clustering methods often focus on directly learning a common representation or a consensus affinity similarity graph from available feature views while ignore the valuable information hidden in the missing views. In this study, we present a novel incomplete multi-view clustering method via adaptive partial graph learning and fusion (APGLF), which can capture the local data structure of both within-view and cross-view. Specifically, we use the available data of each view to learn a corresponding view-specific partial graph, in which the within-view local structure can be well preserved. Then we design a cross-view graph fusion term to learn a consensus complete graph for different views, which can take advantage of the complementary information hidden in the view-specific partial graphs learned from incomplete views. In addition, a rank constraint is imposed on the graph Laplacian matrix of the fused graph to better recover the optimal cluster structure of original data. Therefore, APGLF integrates within-view partial graph learning, cross-view partial graph fusion and cluster structure recovering into a unified framework. Experiments on five incomplete multi-view data sets are conducted to validate the efficacy of APGLF when compared with eight state-of-the-art methods.     

基于双重局部保持的不完整多视角嵌入学习方法

现有的多视角降维方法多数假设数据是完整的,但该假设在实际应用中难以实现。为解决不完整多视角数据降维问题,提出一种新的不完整多视角嵌入学习方法。基于多视角数据的一致性与同一视角下样本间的线性相关性学习一组重构系数,对缺失样本进行线性重构,通过学习所有视角的公共低维嵌入,保持原始空间的局部几何结构。在此基础上,设计一种惩罚参数来度量重构样本的可靠度,从而权衡缺失样本对学习结果的负面影响。实验结果表明,该方法在Yale、ORL和COIL-20数据集上NMI值分别达到65.63%、73.23%和78.27%,较MVL-IV算法分别提升8.37%、16.71%和20.24%。     

mPadal: a joint local-and-global multi-view feature selection method for activity recognition

The selection of multi-view features plays an important role for classifying multi-view data, especially the data with high dimension. In this paper, a novel multi-view feature selection method via joint local pattern-discrimination and global label-relevance analysis (mPadal) is proposed. Different from the previous methods which globally select the multi-view features directly via view-level analysis, the proposed mPadal employs a new joint local-and-global way. In the local selection phase, the pattern-discriminative features will be first selected by considering the local neighbor structure of the most discriminative patterns. In the global selection phase, the features with the topmost label-relevance, which can well separate different classes in the current view, are selected. Finally, the two parts selected are combined to form the final features. Experimental results show that compared with several baseline methods in publicly available activity recognition dataset IXMAS, mPadal performs the best in terms of the highest accuracy, precision, recall and F1 score. Moreover, the features selected by mPadal are highly complementary among views for classification, which is able to improve the classification performance according to previous theoretical studies.     

基于改进结构保持数据降维方法的故障诊断研究

传统基于核主成分分析(Kernel principal component analysis,KPCA)的数据降维方法在提取有效特征信息时只考虑全局结构保持而未考虑样本间的局部近邻结构保持问题,本文提出一种改进全局结构保持算法的特征提取与降维方法.改进的特征提取与降维方法将流形学习中核局部保持投影(Kernel loc...