Article
Computer Science, Artificial Intelligence
Xin Yan, Hongmiao Zhu
Summary: This paper proposes a novel support vector machine model with feature mapping and kernel trick to handle datasets with different distributions. The model improves robustness by pre-selecting training points, and converts the problem into a convex quadratic programming problem solved efficiently by the sequential minimal optimization algorithm. Numerical tests demonstrate the superior performance of the proposed method compared to other classification methods.
KNOWLEDGE-BASED SYSTEMS
(2022)
Article
Computer Science, Interdisciplinary Applications
Xin Yan, Hongmiao Zhu, Jian Luo
Summary: A novel kernel-free Laplacian twin support vector machine method is proposed for semi-supervised classification, which classifies data points into two classes by constructing two nonparallel quadratic surfaces. The method not only saves computational time, but also addresses the issue of computational complexity.
JOURNAL OF COMBINATORIAL OPTIMIZATION
(2021)
Article
Computer Science, Artificial Intelligence
Scindhiya Laxmi, S. K. Gupta, Sumit Kumar
Summary: Intuitionistic fuzzy-based support vector machine is an effective method for multi-category classification problems, which assigns fuzzy score functions to each training point to significantly reduce the impacts of noises and outliers in the dataset.
Article
Computer Science, Artificial Intelligence
F. Marchetti, E. Perracchione
Summary: This article proposes and studies a new efficient tool called Local-to-Global Support Vector Machine (LGSVM) for supervised classification tasks that involve a large number of instances. It constructs a global classifier by gluing together the local SVM contributions via compactly supported weights, which significantly reduces the complexity cost of SVMs for large-scale datasets.
PATTERN RECOGNITION
(2022)
Article
Automation & Control Systems
Guolin Yu, Jun Ma, Chenzhen Xie
Summary: This paper proposes a Hessian scatter regularized twin support vector machine (HSR-TSVM) based on Laplacian regularization. HSR-TSVM can better maintain the local topology of the samples and improve classification performance by utilizing the structural information of samples. Furthermore, a least-squares version of HSR-TSVM called HSR-LSTSVM is proposed to improve computational efficiency.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2023)
Article
Computer Science, Artificial Intelligence
Hongwei Dong, Liming Yang, Xue Wang
Summary: The insufficiency and contamination of supervision information can negatively affect the performance of support vector machines in real-world applications. To address this issue, a novel LK-loss correntropy loss function and LapSVM are introduced for robust semi-supervised classification, showing better generalization performance compared to competing methods.
APPLIED INTELLIGENCE
(2021)
Article
Computer Science, Information Systems
Yangtao Xue, Li Zhang
Summary: The study introduces a novel semi-supervised classification algorithm, LapPVP, which maximizes inter-class distance, minimizes intra-class distance, and preserves data geometry by integrating Laplacian regularization and between-class/within-class scatter optimization. Experiments validate the feasibility and effectiveness of the algorithm.
INFORMATION SCIENCES
(2021)
Article
Computer Science, Hardware & Architecture
Rongyao Hu, Leyuan Zhang, Jian Wei
Summary: A new adaptive LapSVM method is proposed in this paper to achieve semi-supervised learning and improve model accuracy.
Article
Computer Science, Artificial Intelligence
Lan Bai, Xu Chen, Zhen Wang, Yuan-Hai Shao
Summary: This paper proposes a safe intuitionistic fuzzy twin support vector machine (SIFTSVM) for semi-supervised learning. By gradually learning unlabeled samples, it is safer and more precise than learning all of the unlabeled samples simultaneously.
APPLIED SOFT COMPUTING
(2022)
Article
Computer Science, Information Systems
Yuting Sun, Shifei Ding, Lili Guo, Zichen Zhang
Summary: In this paper, a hypergraph regularized semi-supervised support vector machine (HGSVM) algorithm is proposed to exploit the multivariate manifold structure. To accelerate the training process of HGSVM, a fast algorithm based boundary sample selection algorithm, termed fast-HGSVM, is developed. Moreover, two SMOTE-variant techniques and the one-vs-rest strategy are introduced in fast-HGSVM, and two multi-category semi-supervised algorithms called fast-ASHGSVM and fast-KSHGSVM are proposed. The effectiveness of the proposed algorithms is validated through experiments on two moons and UCI datasets.
INFORMATION SCIENCES
(2022)
Article
Computer Science, Information Systems
Sheng Fu, Piao Chen, Zhisheng Ye
Summary: This study proposes a new method called simplex-based proximal MSVM (SPMSVM) as a refinement of the regular multicategory support vector machine (MSVM). The SPMSVM uses a novel family of squared error loss functions and removes the sum-to-zero constraint, resulting in closed-form solutions and reduced computational cost. It can be cast into a weighted regression problem and provides direct probabilistic results, making it more informative and scalable for large-scale applications. Simulations and real examples demonstrate that the SPMSVM is a stable, scalable, and competitive classifier.
IEEE TRANSACTIONS ON INFORMATION THEORY
(2023)
Article
Computer Science, Artificial Intelligence
En Wang, Zi-Yang Wang, Qing Wu
Summary: This paper introduces the semi-supervised support vector machine algorithm for dealing with unlabeled data and its drawbacks, proposing a new class based on the approximation property of Bezier function to enhance generalization and computational efficiency. Experimental results confirm the superiority of the new algorithm in classification accuracy and calculation time.
NEURAL COMPUTING & APPLICATIONS
(2021)
Article
Engineering, Environmental
Zhice Fang, Yi Wang, Hexiang Duan, Ruiqing Niu, Ling Peng
Summary: In this study, different kernel functions of support vector machine (SVM) were compared for landslide susceptibility prediction. The experimental results showed that Laplacian-SVM had the highest prediction performance, and SVMs with RBF kernel were more suitable for this problem compared to SVMs with linear kernel. Furthermore, SVM-based methods demonstrated a higher sensitivity in finding potential landslide areas compared to deep learning methods.
STOCHASTIC ENVIRONMENTAL RESEARCH AND RISK ASSESSMENT
(2022)
Article
Remote Sensing
Christian Geiss, Patrick Aravena Pelizari, Ozan Tuncbilek, Hannes Taubenboeck
Summary: This article introduces two semi-supervised models for remote sensing image classification, which are built upon the framework of Virtual Support Vector Machines (VSVM). These models exhibit the most favorable performance in situations with very limited prior knowledge.
INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION
(2023)
Article
Computer Science, Artificial Intelligence
Scindhiya Laxmi, Sumit Kumar, S. K. Gupta
Summary: Applying machine learning tools to human activity analysis faces two major challenges: the transformation of actions into multiple attributes increases training and testing time significantly, and the presence of noises and outliers in the dataset adds complexity and makes it difficult to implement the activity detection system efficiently. This paper proposes a kernel fuzzy proximal support vector machine as a robust classifier to address these challenges. The proposed method transforms input patterns into a higher-dimensional space and assigns appropriate membership degrees to reduce the impact of noises and outliers, and it achieves computational efficiency by solving a set of linear equations. Simulation results on benchmark problems and numerical results on human activity recognition problems demonstrate the superiority and applicability of the proposed method.
KNOWLEDGE AND INFORMATION SYSTEMS
(2023)
