Unsupervised Feature Selection With Constrained ℓ₂,₀-Norm and Optimized Graph

Nie, Feiping; Dong, Xia; Tian, Lai; Wang, Rong; Li, Xuelong

doi:10.1109/tnnls.2020.3043362

Cited by 38 publications

(8 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…After establishing such a model, we can follow the solution method of Nie et al 25 to solve Equation (13). we can obtain an index vector group

boldq

of selected elements, which is derived from the

k

maximal diagonal elements of

{false(boldA false)}^{- 1} boldB

.…”

Section: The Methodologymentioning

confidence: 99%

“…Still, according to the calculation result of Equation ( 14), its convergence positively correlates with the change of U. For other variables, according to the convergence analysis of Nie et al 25 and Wang et al, 26 the value of the objective function will decrease or remain unchanged at each step of optimizing a variable. Therefore, our alternating update rule can let the objective function value fall monotonically and guarantee the objective function convergence.…”

Section: Complexity and Convergence Analysismentioning

confidence: 98%

See 1 more Smart Citation

Magnetic tile fault detection of high‐voltage electric machines: A multiview feature selection method via consistent soft label

Chen,

Yang,

Liu

et al. 2023

Circuit Theory & Apps

View full text Add to dashboard Cite

Image‐based scheme has attracted wide attention in the fault detection of high‐voltage permanent magnet motors, but it often suffers from the shooting conditions. Multiview feature selection allows multisource information to be fused, which can improve the accuracy and robustness of image detection. Therefore, we propose multiview unsupervised consistency via soft‐label feature selection (MUCSFS). This method constructs consistent pseudo labels through soft labels from clustering affinity of each view sample and builds the model by integrating selection constraints into the graph model. This model filters the fault data set to obtain the feature subset, which is used for clustering. We verify the method's effectiveness by simulating multiview data. The fault clustering experiment on the high‐voltage motors' magnetic tile fault data set confirm that our method can effectively cluster the fault categories.

show abstract

“…After establishing such a model, we can follow the solution method of Nie et al 25 to solve Equation (13). we can obtain an index vector group

boldq

of selected elements, which is derived from the

k

maximal diagonal elements of

{false(boldA false)}^{- 1} boldB

.…”

Section: The Methodologymentioning

confidence: 99%

Section: Complexity and Convergence Analysismentioning

confidence: 98%

Magnetic tile fault detection of high‐voltage electric machines: A multiview feature selection method via consistent soft label

Chen,

Yang,

Liu

et al. 2023

Circuit Theory & Apps

View full text Add to dashboard Cite

show abstract

“…Nie et al. [23] also suggested an unsupervised feature selection method with a row sparsity constraint and optimized graph. Controllable sparse learning [24] was employed to eliminate noise in the completed sample correlations and feature dependency‐based unsupervised feature selection approach.…”

Section: Introductionmentioning

confidence: 99%

“…Wang et al [22] proposed a robust optimal graph clustering framework where a robust graph is constructed with adaptive neighbors to each data sample. Nie et al [23] also suggested an unsupervised feature selection method with a row sparsity constraint and optimized graph. Controllable sparse learning [24] was employed to eliminate noise in the completed sample correlations and feature dependency-based unsupervised feature selection approach.…”

Section: Introductionmentioning

confidence: 99%

Latent low‐rank representation sparse regression model with symmetric constraint for unsupervised feature selection

Guo

Chen

2023

IET Image Processing

View full text Add to dashboard Cite

Unsupervised feature selection is a dimensionality reduction method and has been widely used as an important and indispensable preprocessing step in many tasks. However, realworld data are not only high-dimensional, but also have intrinsic correlations between data points, which are not fully utilized in feature selection. Furthermore, real-world data usually inevitably contain noise or outliers. In order to select features from data more effectively, a sparse regression model based on latent low-rank representation with the symmetric constraint for unsupervised feature selection is proposed. With the coefficient matrix of non-negative symmetric low-rank representation, an affinity matrix characterized by the correlation relationship between data points is adaptively obtained, which reveals the intrinsic geometric relationship, global structure, and discrimination of data points. A latent representation of all data points obtained from this affinity matrix is employed as a pseudolabel, and feature selection is carried out by sparse linear regression. This method performs feature selection in the learned latent space instead of the original data space. An alternating iteration algorithm is designed to solve the proposed model, and its effectiveness and efficiency are verified on several benchmark data sets.

show abstract

“…However, the above three algorithms just solve the relaxed problem from the original l 2,0 -norm problem, which has a tendency to weaken the performance. Nie et al proposed unsupervised feature selection with constrained l 2,0 -norm and optimized graph (RSOGFS) [28]. RSOGFS tackles the l 2,0 -norm problem directly so as to choose the needed features at a time instead of one by one.…”

Section: Introductionmentioning

confidence: 99%

Unsupervised feature selection via self-paced learning and low-redundant regularization

Li¹,

Chen²,

Li³

et al. 2021

Preprint

View full text Add to dashboard Cite

Much more attention has been paid to unsupervised feature selection nowadays due to the emergence of massive unlabeled data. The distribution of samples and the latent effect of training a learning method using samples in more effective order need to be considered so as to improve the robustness of the method. Self-paced learning is an effective method considering the training order of samples. In this study, an unsupervised feature selection is proposed by integrating the framework of self-paced learning and subspace learning. Moreover, the local manifold structure is preserved and the redundancy of features is constrained by two regularization terms. L 2,1/2 -norm is applied to the projection matrix, which aims to retain discriminative features and further alleviate the effect of noise in the data. Then, an iterative method is presented to solve the optimization problem. The convergence of the method is proved theoretically and experimentally. The proposed method is compared with other state of the art algorithms on nine realworld datasets. The experimental results show that the proposed method can improve the performance of clustering methods and outperform other compared algorithms.

show abstract

Unsupervised Feature Selection With Constrained ℓ₂,₀-Norm and Optimized Graph

Cited by 38 publications

References 28 publications

Magnetic tile fault detection of high‐voltage electric machines: A multiview feature selection method via consistent soft label

Magnetic tile fault detection of high‐voltage electric machines: A multiview feature selection method via consistent soft label

Latent low‐rank representation sparse regression model with symmetric constraint for unsupervised feature selection

Unsupervised feature selection via self-paced learning and low-redundant regularization

Contact Info

Product

Resources

About