Article
Computer Science, Information Systems
Weichan Zhong, Xiaojun Chen, Feiping Nie, Joshua Zhexue Huang
Summary: In this paper, a novel adaptive discriminant analysis method SADA is proposed for semi-supervised feature selection, which can effectively learn the similarity matrix and projection matrix during the process. Experimental results demonstrate the superior performance of SADA compared to other semi-supervised feature selection methods.
INFORMATION SCIENCES
(2021)
Article
Computer Science, Information Systems
Chenglong Zhang, Bingbing Jiang, Zidong Wang, Jie Yang, Yangfeng Lu, Xingyu Wu, Weiguo Sheng
Summary: In this paper, an efficient multi-view feature selection method (EMSFS) is proposed to address the issues in multi-view semi-supervised feature selection. EMSFS combines graph learning, label propagation, and multi-view feature selection within a unified framework. The method can adaptively learn a graph and exploit the similarity structure to enhance the reliability of the graph. It also achieves high computational efficiency.
INFORMATION SCIENCES
(2023)
Article
Computer Science, Information Systems
Jingliu Lai, Hongmei Chen, Tianrui Li, Xiaoling Yang
Summary: In this study, a novel semi-supervised sparse feature selection framework is proposed, which improves the quality of the similarity matrix through adaptive graph learning and alleviates the negative influence of redundant features through redundancy minimization regularization.
INFORMATION SCIENCES
(2022)
Article
Computer Science, Artificial Intelligence
Qingqing Pang, Li Zhang
Summary: This paper proposes a Recursive Feature Retention (RFR) method based on NDI and FILS methods, aiming to determine an optimal feature subset with high discriminant ability and ability to maintain local data structure. RFR iteratively selects features by compromising scores to achieve better performance compared with state-of-the-art semi-supervised methods.
INTERNATIONAL JOURNAL OF MACHINE LEARNING AND CYBERNETICS
(2021)
Article
Computer Science, Interdisciplinary Applications
Tue Boesen, Eldad Haber, G. Michael Hoversten
Summary: A new graph-Laplacian based semi-supervised clustering method is proposed in this research, which is capable of handling large datasets and yielding satisfactory results in oil prospectivity analysis.
COMPUTERS & GEOSCIENCES
(2021)
Article
Automation & Control Systems
Xuemin Tan, Jun Qi, John Q. Gan, Jianglin Zhang, Chao Guo, Fu Wan, Ke Wang
Summary: This paper proposes a novel multi-filter semi-supervised feature selection method for optimizing DGA features and improving fault diagnosis accuracy using limited labeled DGA samples and more unlabeled DGA data. The method ranks input DGA features based on five different evaluation criteria and aggregates feature ranks by multiple filters to form a lower-dimensional candidate feature subset. The results show that the proposed method works well in optimizing DGA features and significantly improving fault diagnosis accuracy, and the robustness of the selection of optimal feature subset is validated.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2023)
Article
Computer Science, Artificial Intelligence
Jingliu Lai, Hongmei Chen, Weiyi Li, Tianrui Li, Jihong Wan
Summary: In this study, a novel semi-supervised feature selection model ASLCGLFS is proposed, which combines label information to extend adaptive graph learning for further improving the quality of the similarity matrix. Additionally, adaptive structure learning is introduced to consider global structure and facilitate feature selection.
KNOWLEDGE-BASED SYSTEMS
(2022)
Article
Computer Science, Information Systems
Sichao Fu, Weifeng Liu, Kai Zhang, Yicong Zhou, Dapeng Tao
Summary: This paper introduces the graph p-Laplacian convolutional networks (GpLCN) to better extract sample features and improve classification performance by utilizing graph p-Laplacian matrix.
INFORMATION SCIENCES
(2021)
Article
Mathematics, Applied
Andrea L. Bertozzi, Bamdad Hosseini, Hao Li, Kevin Miller, Andrew M. Stuart
Summary: This paper explores the consistency of graph-based semi-supervised regression in classification tasks by analyzing the impact of label noise and inherent clustering in the graph on the posterior measure.
Article
Computer Science, Artificial Intelligence
Guolin Zhang, Zehui Hu, Guoqiu Wen, Junbo Ma, Xiaofeng Zhu
Summary: The traditional graph convolutional network(GCN) and its variants usually only propagate node information through the given topology, ignoring some correlative feature information between nodes. To address this issue, a novel model named Dynamic Graph Convolutional Networks by Semi-Supervised Contrastive Learning (DGSCL) is proposed in this paper. It constructs a dynamic feature graph from the input node features and fuses it with the given topology using co-attention modules for more informative node embeddings.
PATTERN RECOGNITION
(2023)
Article
Computer Science, Artificial Intelligence
Daxin Huang, Junzheng Jiang, Fang Zhou, Shan Ouyang
Summary: In this paper, a novel distributed algorithm is proposed for solving the graph semi-supervised learning problem, which only requires neighboring information and the difference between solutions is upper bounded compared to the centralized approach. The effectiveness of the proposed distributed algorithm is verified through numerical results on synthetic and real-world datasets.
PATTERN RECOGNITION LETTERS
(2021)
Article
Computer Science, Artificial Intelligence
Jiaqi Bao, Mineichi Kudo, Keigo Kimura, Lu Sun
Summary: A novel approach called robust embedding regression (RER) for semi-supervised learning is proposed in this paper, which constructs a more robust and accurate graph by inheriting the advantages of existing techniques and introduces low-rank representation to reduce the negative influence of redundant features and noises. Experimental results demonstrate that RER outperforms existing methods in classification and clustering performance.
PATTERN RECOGNITION
(2024)
Article
Computer Science, Artificial Intelligence
Xinmin Tao, Yixuan Bao, Xiaohan Zhang, Tian Liang, Lin Qi, Zhiting Fan, Shan Huang
Summary: A semi-supervised kernel local Fisher discriminant analysis algorithm based on density peak clustering pseudo-labels (SDPCKLFDA) is proposed to effectively use both labeled and unlabeled data for learning. The algorithm generates pseudo cluster labels using the density peak clustering algorithm and constructs regularization strategies based on these pseudo-labels. The optimal projection vector is obtained by solving the objective function of the local Fisher discriminant analysis. The algorithm improves the discriminant performance of extracted features and is suitable for multimodal and noisy data.
NEURAL COMPUTING & APPLICATIONS
(2022)
Article
Computer Science, Artificial Intelligence
Razieh Sheikhpour
Summary: Feature selection is widely used in machine learning applications to select relevant features from data sets. Recently, there has been considerable research interest in semi-supervised sparse feature selection based on graph Laplacian, which uses the correlation between features. This paper proposes a spline regression-based framework for semi-supervised sparse feature selection, which uses mixed convex and non-convex t2,p-norm regularization to select relevant features and considers feature correlation. The framework retains the geometry structure of labeled and unlabeled data using local spline regression and encodes the data distribution. A unified iterative algorithm is presented to solve the framework, and its convergence is theoretically and experimentally proved. Experiments on several data sets demonstrate the effectiveness of the framework in selecting the most relevant and discriminative features.
KNOWLEDGE-BASED SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Sitao Lv, Shengfei Shi, Hongzhi Wang, Feng Li
Summary: The paper proposes a multi-label feature selection method SFAM, which combines global structure learning and manifold learning to address the challenges of high-dimensional multi-label data. This method achieves excellent performance through the design of an efficient iterative algorithm.
KNOWLEDGE-BASED SYSTEMS
(2021)
Article
Computer Science, Information Systems
Razieh Sheikhpour, Mehdi Agha Sarram, Elnaz Sheikhpour
INFORMATION SCIENCES
(2018)
Article
Telecommunications
Azam Jannesari, Mehdi Agha Sarram, Razieh Sheikhpour
WIRELESS PERSONAL COMMUNICATIONS
(2020)
Article
Automation & Control Systems
Razieh Sheikhpour, Sajjad Gharaghani, Elmira Nazarshodeh
CHEMOMETRICS AND INTELLIGENT LABORATORY SYSTEMS
(2020)
Article
Computer Science, Information Systems
Razieh Sheikhpour, Mehdi Agha Sarram, Sajjad Gharaghani, Mohammad Ali Zare Chahooki
INFORMATION SCIENCES
(2020)
Article
Automation & Control Systems
Mohammad Morovvati Sharifabad, Razieh Sheikhpour, Sajjad Gharaghani
Summary: De novo drug discovery is a costly and time-consuming process. Repositioning existing drugs for new applications can reduce the time and cost of finding new drugs. Predicting drug-target interactions (DTIs) can facilitate drug repositioning, but there are challenges due to the diversity of drug descriptors and protein features, as well as the lack of experimentally-confirmed non-interacting drug-target pairs as negative samples. This study presents a modified algorithm for extracting balanced negative samples and a semi-supervised feature selection method, which outperform other methods on benchmark DTI datasets.
CHEMOMETRICS AND INTELLIGENT LABORATORY SYSTEMS
(2022)
Article
Biochemistry & Molecular Biology
Zahra Bastami, Razieh Sheikhpour, Parvin Razzaghi, Ali Ramazani, Sajjad Gharaghani
Summary: Caspases are important enzymes involved in inflammation and cell death processes. This study used Proteochemometrics Modeling to summarize and predict the interactions between caspases and ligands. The ensemble model showed superior performance compared to other models.
MOLECULAR DIVERSITY
(2023)
Article
Pharmacology & Pharmacy
Mohammad Morovvati Sharifabad, Razieh Sheikhpour, Sajjad Gharaghani
Summary: This study proposes a reliable algorithm for selecting negative samples in drug-target interaction prediction, which demonstrates superior performance and highlights the significant improvement in learning process performance by correctly selecting negative samples.
JOURNAL OF PHARMACOLOGICAL AND TOXICOLOGICAL METHODS
(2022)
Article
Soil Science
Ruhollah Taghizadeh-Mehrjardi, Razieh Sheikhpour, Mojtaba Zeraatpisheh, Alireza Amirian-Chakan, Norair Toomanian, Ruth Kerry, Thomas Scholten
Summary: Digital soil mapping can be used to predict soils at unvisited sites, but problems arise when predictions are needed in areas without any soil observations. A new semi-supervised learning approach was found to outperform supervised learning in extrapolating soil classes in target areas, resulting in higher accuracy and lower uncertainty.
Article
Computer Science, Software Engineering
Roohallah Fazli, Hadi Owlia, Razieh Sheikhpour
Summary: A robust algorithm for source number estimation based on the formation of the Hankel covariance matrix is presented. The proposed algorithm can handle both non-coherent and fully coherent sources, and it outperforms competing methods in numerical simulations.
INTERNATIONAL JOURNAL OF WAVELETS MULTIRESOLUTION AND INFORMATION PROCESSING
(2023)
Article
Computer Science, Artificial Intelligence
Razieh Sheikhpour
Summary: Feature selection is widely used in machine learning applications to select relevant features from data sets. Recently, there has been considerable research interest in semi-supervised sparse feature selection based on graph Laplacian, which uses the correlation between features. This paper proposes a spline regression-based framework for semi-supervised sparse feature selection, which uses mixed convex and non-convex t2,p-norm regularization to select relevant features and considers feature correlation. The framework retains the geometry structure of labeled and unlabeled data using local spline regression and encodes the data distribution. A unified iterative algorithm is presented to solve the framework, and its convergence is theoretically and experimentally proved. Experiments on several data sets demonstrate the effectiveness of the framework in selecting the most relevant and discriminative features.
KNOWLEDGE-BASED SYSTEMS
(2023)
Article
Pediatrics
Sanaz Mehrabani, Morteza Zangeneh Soroush, Negin Kheiri, Razieh Sheikhpour, Mahshid Bahrami
Summary: This study aimed to predict blood cancer using leukemia gene expression data and a robust l2,p-norm sparsity-based gene selection method. The results showed that this method can correctly classify all samples of acute myeloid leukemia (AML) and lymphoblastic leukemia (ALL), and identified seven important genes, with PRTN3 gene being the most important. This method can be useful for predicting leukemia and examining the expression levels of related genes.
IRANIAN JOURNAL OF PEDIATRIC HEMATOLOGY AND ONCOLOGY
(2023)
Article
Computer Science, Artificial Intelligence
Razieh Sheikhpour, Kamal Berahmand, Saman Forouzandeh
Summary: Feature selection aims to eliminate redundant features and choose informative ones. Semi-supervised feature selection becomes important as it utilizes labeled and unlabeled data. We propose two frameworks, one based on Hessian matrix and the other on Hessian-Laplacian combination, for semi-supervised feature selection. Our frameworks utilize regularization and constraint techniques to select informative features and maintain the topological structure of data. Experimental results demonstrate the effectiveness of our frameworks in selecting informative features.
KNOWLEDGE-BASED SYSTEMS
(2023)
Article
Pediatrics
Razieh Sheikhpour, Roohallah Fazli, Sanaz Mehrabani
Summary: This study identified important genes for the diagnosis of acute myeloid and lymphoblastic leukemia using microarray data and a sparse feature selection method. The results showed that AML and ALL can be accurately diagnosed with high accuracy using machine learning methods. The investigation of selected genes in this study may be helpful for the diagnosis of ALL and AML.
IRANIAN JOURNAL OF PEDIATRIC HEMATOLOGY AND ONCOLOGY
(2021)
Article
Engineering, Multidisciplinary
Mohammad Momeny, Ali Mohammad Latif, Mehdi Agha Sarram, Razieh Sheikhpour, Yu Dong Zhang
Summary: In this paper, a Noise-Robust Convolutional Neural Network (NR-CNN) is proposed to classify noisy images without preprocessing, by adding a noise map layer and an adaptive resize layer, and considering noise in different components of the network. The proposed NR-CNN improves the classification performance of noisy images and network training speed.
RESULTS IN ENGINEERING
(2021)