Jie Li
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Research Interests
Jie Li has a total of 9 co-authors. Here are the top co-authors by number of shared publications:
Xuzhu Shen
5 shared publicationsChaoqian Li
4 shared publications- Chaoqian Li
School of Mathematics and Statistics, Yunan University, Kunming, China
2 shared publications - Yaotang Li
School of Mathematics and Statistics, Yunnan University, Kunming, China
2 shared publications - Chaogang Shao
Affiliation College of Life Sciences, Huzhou University, Huzhou, Zhejiang, P.R. China
1 shared publication - Xiaoxia Ma
College of Pharmaceutical Science Zhejiang Univ. of Technology Hangzhou 310014 China
1 shared publication
Latest Publications
- 13 August 2026
- Published by Elsevier in Pattern Recognition
- Vol. 180
High-order label propagation-guided nonnegative matrix factorization for semi-supervised clustering
- 13 August 2026
- Published by Elsevier in Pattern Recognition
- Vol. 180
High-order label propagation-guided nonnegative matrix factorization for semi-supervised clustering
- Open Access
- 22 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 78907-78919
Locally Constrained Non-Negative Matrix Factorization for Data Clustering
Semi-supervised non-negative matrix factorization (NMF) has been widely used for clustering high-dimensional data, because a small amount of label information can effectively improve clustering performance. Nevertheless, most existing semi-supervisedNMFmethods still face two important challenges: they require carefully tuned hyperparameters to control the effect of supervision, and they cannot directly produce clustering assignments from the learned factorization results. In this paper, we propose a novel semi-supervised NMF framework, called Locally Constrained NMF (LCNMF), to address both issues simultaneously. Different from existing methods that impose global penalty terms or fit predefined label codes, LCNMF injects label information by encoding locally inconsistent elements in the low-dimensional representations of labeled samples as zero, thereby introducing an explicit local structural constraint into NMF. This design eliminates the need for additional hyperparameter tuning and allows the learned representation to carry more direct clustering semantics. Furthermore, LCNMF integrates dimensionality reduction and clustering assignment into a unified framework, so that the final cluster labels can be obtained directly without an extra clustering step.We further formulate two variants of the proposed model and derive the corresponding optimization algorithms. Extensive experiments on several real-world datasets show that LCNMF consistently achieves effective and competitive clustering performance compared with state-of-the-art semi-supervised clustering methods. The code can be obtained at https://github.com/ljisxz/LCNMF.
- 22 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 78907-78919
Open AccessLocally Constrained Non-Negative Matrix Factorization for Data Clustering
Semi-supervised non-negative matrix factorization (NMF) has been widely used for clustering high-dimensional data, because a small amount of label information can effectively improve clustering performance. Nevertheless, most existing semi-supervisedNMFmethods still face two important challenges: they require carefully tuned hyperparameters to control the effect of supervision, and they cannot directly produce clustering assignments from the learned factorization results. In this paper, we propose a novel semi-supervised NMF framework, called Locally Constrained NMF (LCNMF), to address both issues simultaneously. Different from existing methods that impose global penalty terms or fit predefined label codes, LCNMF injects label information by encoding locally inconsistent elements in the low-dimensional representations of labeled samples as zero, thereby introducing an explicit local structural constraint into NMF. This design eliminates the need for additional hyperparameter tuning and allows the learned representation to carry more direct clustering semantics. Furthermore, LCNMF integrates dimensionality reduction and clustering assignment into a unified framework, so that the final cluster labels can be obtained directly without an extra clustering step.We further formulate two variants of the proposed model and derive the corresponding optimization algorithms. Extensive experiments on several real-world datasets show that LCNMF consistently achieves effective and competitive clustering performance compared with state-of-the-art semi-supervised clustering methods. The code can be obtained at https://github.com/ljisxz/LCNMF.
- Open Access
- 12 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 73189-73200
Improved Graph-Regularized Discriminative Nonnegative Matrix Factorization for Semi-Supervised Clustering
Nonnegative matrix factorization (NMF) is an effective dimensionality reduction and representation learning technique that captures the intrinsic structure of nonnegative data by learning low-dimensional, parts-based representations. By incorporating a small amount of label information, semi-supervised NMF can further enhance the discriminative power of the learned representations for clustering tasks. In this paper, we propose an improved graph-regularized discriminative nonnegative matrix factorization model, termed IGDNMF. The proposed model preserves the local manifold structure through graph Laplacian regularization while introducing discriminative constraints that scale and rotate the representations of labeled samples, encouraging them to align along class-specific axes in the learned representation space and thereby improving class separability. Extensive clustering experiments on five real-world datasets demonstrate that IGDNMF consistently outperforms several state-of-the-art semi-supervised NMF methods.
- 12 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 73189-73200
Open AccessImproved Graph-Regularized Discriminative Nonnegative Matrix Factorization for Semi-Supervised Clustering
Nonnegative matrix factorization (NMF) is an effective dimensionality reduction and representation learning technique that captures the intrinsic structure of nonnegative data by learning low-dimensional, parts-based representations. By incorporating a small amount of label information, semi-supervised NMF can further enhance the discriminative power of the learned representations for clustering tasks. In this paper, we propose an improved graph-regularized discriminative nonnegative matrix factorization model, termed IGDNMF. The proposed model preserves the local manifold structure through graph Laplacian regularization while introducing discriminative constraints that scale and rotate the representations of labeled samples, encouraging them to align along class-specific axes in the learned representation space and thereby improving class separability. Extensive clustering experiments on five real-world datasets demonstrate that IGDNMF consistently outperforms several state-of-the-art semi-supervised NMF methods.
- 10 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems for Video Technology
- Vol. 36 (6) , 8342-8355
Multi-View Clustering via Graph Regularized and One-Hot-Constrained Symmetric Nonnegative Matrix Factorization
The semi-supervised multiview clustering (MVC) methods based on non-negative matrix factorization (NMF) have attracted considerable attention due to their ability to utilize partial supervision information and significantly enhance clustering performance. Still, they often face challenges such as high computational costs and difficulties in parameter tuning. We propose a semi-supervised MVC method called Graph-Regularized and One-Hot-Constrained Symmetric NMF (MGOCS) to avoid these issues relatively. The key advantage of MGOCS includes the introduction of a parameter-free semi-supervised strategy that encodes partial label information as one-hot vectors during decomposition, the use of a shared consistency indicator matrix directly as the clustering assignment matrix for all views, and the incorporation of graph regularization terms for each view. These features effectively balance model complexity with information utilization. An iterative optimization algorithm is developed to solve MGOCS, accompanied by complexity and convergence analysis. Clustering experiments conducted on six multi-view datasets demonstrate that MGOCS outperforms various state-of-the-art methods in terms of clustering performance. The code is available at https://github.com/ljisxz/MGOCS.
- 10 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems for Video Technology
- Vol. 36 (6) , 8342-8355
Multi-View Clustering via Graph Regularized and One-Hot-Constrained Symmetric Nonnegative Matrix Factorization
The semi-supervised multiview clustering (MVC) methods based on non-negative matrix factorization (NMF) have attracted considerable attention due to their ability to utilize partial supervision information and significantly enhance clustering performance. Still, they often face challenges such as high computational costs and difficulties in parameter tuning. We propose a semi-supervised MVC method called Graph-Regularized and One-Hot-Constrained Symmetric NMF (MGOCS) to avoid these issues relatively. The key advantage of MGOCS includes the introduction of a parameter-free semi-supervised strategy that encodes partial label information as one-hot vectors during decomposition, the use of a shared consistency indicator matrix directly as the clustering assignment matrix for all views, and the incorporation of graph regularization terms for each view. These features effectively balance model complexity with information utilization. An iterative optimization algorithm is developed to solve MGOCS, accompanied by complexity and convergence analysis. Clustering experiments conducted on six multi-view datasets demonstrate that MGOCS outperforms various state-of-the-art methods in terms of clustering performance. The code is available at https://github.com/ljisxz/MGOCS.
- 22 May 2025
- Published by Elsevier in Engineering Applications of Artificial Intelligence
- Vol. 156
Elements discriminative non-negative matrix factorization for data clustering
- 22 May 2025
- Published by Elsevier in Engineering Applications of Artificial Intelligence
- Vol. 156
Elements discriminative non-negative matrix factorization for data clustering
- 7 February 2025
- Published by Elsevier in Pattern Recognition
- Vol. 162
One-hot constrained symmetric nonnegative matrix factorization for image clustering
- 7 February 2025
- Published by Elsevier in Pattern Recognition
- Vol. 162
One-hot constrained symmetric nonnegative matrix factorization for image clustering
- Open Access
- 27 December 2023
- Published by MDPI AG in Mathematics
- Vol. 12 (1) , 96
Dual-Graph-Regularization Constrained Nonnegative Matrix Factorization with Label Discrimination for Data Clustering
NONNEGATIVE matrix factorization (NMF) is an effective technique for dimensionality reduction of high-dimensional data for tasks such as machine learning and data visualization. However, for practical clustering tasks, traditional NMF ignores the manifold information of both the data space and feature space, as well as the discriminative information of the data. In this paper, we propose a semisupervised NMF called dual-graph-regularization-constrained nonnegative matrix factorization with label discrimination (DCNMFLD). DCNMFLD combines dual graph regularization and prior label information as additional constraints, making full use of the intrinsic geometric and discriminative structures of the data, and can efficiently enhance the discriminative and exclusionary nature of clustering and improve the clustering performance. The evaluation of the clustering experimental results on four benchmark datasets demonstrates the effectiveness of our new algorithm.
- 27 December 2023
- Published by MDPI AG in Mathematics
- Vol. 12 (1) , 96
Open AccessDual-Graph-Regularization Constrained Nonnegative Matrix Factorization with Label Discrimination for Data Clustering
NONNEGATIVE matrix factorization (NMF) is an effective technique for dimensionality reduction of high-dimensional data for tasks such as machine learning and data visualization. However, for practical clustering tasks, traditional NMF ignores the manifold information of both the data space and feature space, as well as the discriminative information of the data. In this paper, we propose a semisupervised NMF called dual-graph-regularization-constrained nonnegative matrix factorization with label discrimination (DCNMFLD). DCNMFLD combines dual graph regularization and prior label information as additional constraints, making full use of the intrinsic geometric and discriminative structures of the data, and can efficiently enhance the discriminative and exclusionary nature of clustering and improve the clustering performance. The evaluation of the clustering experimental results on four benchmark datasets demonstrates the effectiveness of our new algorithm.
- Open Access
Detect feature edges for diagnosis of bacterial vaginosis
One of the most common diseases among women of reproductive age is bacterial vaginosis (BV). However, the etiology of BV remains unknown. In this study, we modeled the temporal sample of the vaginal microbiome as a network and investigated the relationship between the network edges and BV. Furthermore, we used feature selection algorithms including decision tree (DT) and ReliefF (RF) to select the network feature edges associated with BV and subsequently validated these feature edges through logistic regression (LR) and support vector machine (SVM). The results show that: machine learning can distinguish vaginal community states (BV, ABV, SBV, and HEA) based on a few feature edges; selecting the top five feature edges of importance can achieve the best accuracy for the feature selection and classification model; the feature edges selected by DT outperform those selected by RF in terms of classification algorithm LR and SVM, and LR with DT feature edges is more suitable for diagnosing BV; two feature selection algorithms exhibit differences in the importance of ranking of edges; the feature edges selected by DT and RF cannot construct sub-network associated with BV. In short, the feature edges selected by our method can serve as indicators for personalized diagnosis of BV and aid in the clarification of a more mechanistic interpretation of its etiology.
Open AccessDetect feature edges for diagnosis of bacterial vaginosis
One of the most common diseases among women of reproductive age is bacterial vaginosis (BV). However, the etiology of BV remains unknown. In this study, we modeled the temporal sample of the vaginal microbiome as a network and investigated the relationship between the network edges and BV. Furthermore, we used feature selection algorithms including decision tree (DT) and ReliefF (RF) to select the network feature edges associated with BV and subsequently validated these feature edges through logistic regression (LR) and support vector machine (SVM). The results show that: machine learning can distinguish vaginal community states (BV, ABV, SBV, and HEA) based on a few feature edges; selecting the top five feature edges of importance can achieve the best accuracy for the feature selection and classification model; the feature edges selected by DT outperform those selected by RF in terms of classification algorithm LR and SVM, and LR with DT feature edges is more suitable for diagnosing BV; two feature selection algorithms exhibit differences in the importance of ranking of edges; the feature edges selected by DT and RF cannot construct sub-network associated with BV. In short, the feature edges selected by our method can serve as indicators for personalized diagnosis of BV and aid in the clarification of a more mechanistic interpretation of its etiology.
- 1 February 2022
- Published by Elsevier in Journal of Plant Physiology
- Vol. 270, 153636
Creating and maintaining a high-confidence microRNA repository for crop research: A brief review and re-examination of the current crop microRNA registries
- 1 February 2022
- Published by Elsevier in Journal of Plant Physiology
- Vol. 270, 153636
Creating and maintaining a high-confidence microRNA repository for crop research: A brief review and re-examination of the current crop microRNA registries