Article
Management
Asuncion Jimenez-Cordero, Juan Miguel Morales, Salvador Pineda
Summary: Feature selection has become a challenging issue in machine learning, particularly in classification problems. Support Vector Machine is a widely used technique in classification tasks, with various methodologies proposed for selecting the most relevant features in SVM. The authors introduce an embedded feature selection method based on a min-max optimization problem to balance model complexity and classification accuracy, showcasing efficiency and usefulness in benchmark datasets.
EUROPEAN JOURNAL OF OPERATIONAL RESEARCH
(2021)
Article
Computer Science, Information Systems
Farid Yuli Martin Adiyatma, Dwi Joko Suroso, Panarat Cherntanomwong
Summary: This paper introduces two improved Min-Max algorithms, TLB-MM and WC-TLB-MM, to solve the issue of obstructed indoor localization accuracy. The effectiveness of these methods is validated through experiments conducted in a laboratory room.
Article
Computer Science, Artificial Intelligence
Zongmin Liu, Yitian Xu
Summary: In this paper, a novel multi-task nonparallel support vector machine (MTNPSVM) is proposed, which effectively avoids matrix inversion operation and takes full advantage of the kernel trick by introducing epsilon-insensitive loss instead of square loss. The alternating direction method of multipliers (ADMM) is employed to improve computational efficiency, and the properties and sensitivity of the model parameters are further explored.
APPLIED SOFT COMPUTING
(2022)
Article
Computer Science, Information Systems
Jinseong Park, Yujin Choi, Junyoung Byun, Jaewook Lee, Saerom Park
Summary: In this paper, a multi-class classification method using kernel supports and a dynamical system under differential privacy is proposed. For small datasets, kernel methods, such as kernel support vector machines (SVMs), show good generalization performance with high-dimensional feature mapping. However, kernel SVMs have a fundamental weakness in achieving differential privacy because they construct decision functions based on a subset of the training data called support vectors. To address these limitations, a two-phase classification algorithm based on support vector data description (SVDD) is developed. It generates a differentially private SVDD (DP-SVDD) by perturbing the sphere center in a high-dimensional feature space and partitions the input space using a dynamical system for classification.
INFORMATION SCIENCES
(2023)
Article
Biochemistry & Molecular Biology
Vineet Thumuluri, Jose Juan Almagro Armenteros, Alexander Rosenberg Johansen, Henrik Nielsen, Ole Winther
Summary: This article introduces an upgraded version of the DeepLoc tool for predicting protein subcellular localization. By using a pre-trained protein language model and providing features such as attention outputs and sorting signal prediction, DeepLoc 2.0 achieves state-of-the-art performance and interpretability.
NUCLEIC ACIDS RESEARCH
(2022)
Article
Computer Science, Artificial Intelligence
Yinan Guo, Zirui Zhang, Fengzhen Tang
Summary: Feature selection is important in machine learning to reduce complexity and simplify interpretation. A novel non-linear method proposed in this paper uses kernelized multi-class support vector machines and fast recursive feature elimination to select features that work well for all classes, resulting in lower computational time complexity.
PATTERN RECOGNITION
(2021)
Article
Computer Science, Information Systems
Zhao-Yue Zhang, Zi-Jie Sun, Yu-He Yang, Hao Lin
Summary: This study presents a support vector machine-based approach that incorporates mutual information algorithm and incremental feature selection strategy to improve the prediction performance of lncRNA subcellular localization.
FRONTIERS OF COMPUTER SCIENCE
(2022)
Review
Biochemical Research Methods
Jun Wang, Marc Horlacher, Lixin Cheng, Ole Winther
Summary: RNA localization is important for spatial translation regulation, and this review discusses its molecular mechanisms, experimental techniques, and machine learning-based prediction tools. The three main molecular mechanisms controlling RNA localization to distinct cellular compartments, including directed transport, mRNA degradation protection, and diffusion/local entrapment, are reviewed. Advances in experimental methods provide ample data resources for the design of powerful machine learning models in RNA localization prediction. The review also covers publicly available predictive tools, serving as a guide for users and encouraging the development of more effective prediction models. Lastly, an overview of multimodal learning is presented as a potential new avenue for RNA localization prediction.
BRIEFINGS IN BIOINFORMATICS
(2023)
Article
Engineering, Industrial
Ling Chunyan, Lei Jingzhe, Kuo Way
Summary: This paper proposes an effective method to mitigate computational burden in reliability-based design optimization of modular systems. The method tackles coupling effects of modules using the individual module feasible approach and builds an alternative model for the probabilistic constraint function using Bayesian-inference-based support vector machine. The optimal decision scheme is obtained by solving the formulated conventional RBDO using the alternative model.
RELIABILITY ENGINEERING & SYSTEM SAFETY
(2022)
Article
Computer Science, Artificial Intelligence
Chunling Lou, Xijiong Xie
Summary: Two novel multi-view intuitionistic fuzzy support vector machines with insensitive pinball loss are proposed in this paper, which can handle general multi-view classification problems and be robust to noisy data. The pinball loss is incorporated into the multi-view learning to maximize the quantization distance. Intuitionistic fuzzy score is introduced to assign weights to the multi-view samples to effectively utilize multi-view information.
Article
Computer Science, Artificial Intelligence
Huiru Wang, Jiayi Zhu, Siyuan Zhang
Summary: Multi-view learning aims to utilize different views to complement each other and extract potential information in the data. This paper introduces safe screening rules, SSR-SVM-2K and SSR-MvTwSVM, based on the optimality conditions of SVM-2K and MvTwSVM, respectively. These rules can assign or delete different dual variables in advance to reduce the scale of the optimization problem and improve solution speed. Additionally, a sequence screening rule is proposed to accelerate parameter optimization process, and its properties are analyzed.
Article
Computer Science, Artificial Intelligence
Ran An, Yitian Xu, Xuhua Liu
Summary: TSVM is suitable for STL problems, while MTL explores shared information between multiple tasks for better classification. The proposed rough MT-v-TSVM assigns different penalties to misclassified samples based on their positions, combining the advantages of rough v-TSVM and preserving the individuality of tasks.
APPLIED SOFT COMPUTING
(2021)
Article
Mathematical & Computational Biology
Yongyin Han, Maolin Liu, Zhixiao Wang
Summary: This paper presents a novel method for key protein identification that integrates protein complex information with multiple biological features. The proposed method achieves higher accuracy in identifying key proteins compared to classical methods and exhibits robustness across diverse protein-protein interaction networks.
MATHEMATICAL BIOSCIENCES AND ENGINEERING
(2023)
Article
Automation & Control Systems
Yunhao Zhang, Jiajun Yu, Xinyi Dong, Ping Zhong
Summary: Two novel multi-task support vector machines with pinball loss are proposed for binary classification, which maximize the quantile distance for each task to be less sensitive to noise and more stable for re-sampling. The models can achieve better performance by choosing suitable combinations of kernel functions for different tasks, and also include the multi-task SVM with hinge loss as their special cases. Extensive experiments on multi-task datasets validate the validity of the proposed models.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2021)
Article
Biochemical Research Methods
Yue Bi, Fuyi Li, Xudong Guo, Zhikang Wang, Tong Pan, Yuming Guo, Geoffrey Webb, Jianhua Yao, Cangzhi Jia, Jiangning Song
Summary: Subcellular localization of mRNAs is crucial for the spatial regulation of gene activity. In this study, a novel predictor called Clarion was proposed for mRNA subcellular localization prediction. Clarion achieved competitive predictive performance, outperforming state-of-the-art methods, and can accurately predict the localization of mRNAs in different subcellular regions.
BRIEFINGS IN BIOINFORMATICS
(2022)
Article
Engineering, Biomedical
Xun Wu, Wei-Long Zheng, Ziyi Li, Bao-Liang Lu
Summary: This study proposes a novel algorithm for selecting emotion-relevant critical subnetworks and investigates three EEG functional connectivity network features. The results show that these EEG connectivity features achieve high classification accuracy in emotion recognition.
JOURNAL OF NEURAL ENGINEERING
(2022)
Article
Automation & Control Systems
Wei Wu, Wei Sun, Q. M. Jonathan Wu, Yimin Yang, Hui Zhang, Wei-Long Zheng, Bao-Liang Lu
Summary: This article discusses the phenomenon of increasing accidents caused by reduced vigilance and proposes a method based on a multimodal regression network with feature fusion to improve accuracy and efficiency.
IEEE TRANSACTIONS ON CYBERNETICS
(2022)
Review
Computer Science, Artificial Intelligence
Dongrui Wu, Yifan Xu, Bao-Liang Lu
Summary: This article reviews journal publications on transfer learning approaches in EEG-based BCIs since 2016. TL methods applied to different paradigms and applications, such as motor imagery and event-related potentials, are reviewed in the context of cross-subject/session and cross-device/task settings. Observations and conclusions provide insights for future research directions.
IEEE TRANSACTIONS ON COGNITIVE AND DEVELOPMENTAL SYSTEMS
(2022)
Article
Engineering, Electrical & Electronic
Xing Li, Fangyao Shen, Yong Peng, Wanzeng Kong, Bao-Liang Lu
Summary: Recently, there has been increasing interest in research on emotion recognition based on electroencephalogram (EEG). The properties of weak, non-stationary, multi-rhythm and multi-channel EEG data easily lead to different contributions of extracted EEG samples and features in recognizing emotional states. However, existing studies have either neglected the importance of both samples and features or only considered one of them. In this study, a new model called sJSFE (semi-supervised Joint Sample and Feature importance Evaluation) is proposed to quantitatively measure the importance of samples and features using self-paced learning and feature self-weighting, respectively. Experimental results on the SEED-IV dataset demonstrate that mining both sample and feature importance greatly improves the performance of emotion recognition. The average accuracy obtained by sJSFE across the three cross-session recognition tasks is 82.45%, which is significantly higher than the results of traditional models. Furthermore, the feature importance vector reveals that the Gamma frequency band contributes the most, and specific brain regions are correlated with emotion recognition.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS II-EXPRESS BRIEFS
(2022)
Article
Engineering, Biomedical
Yong Peng, Honggang Liu, Junhua Li, Jun Huang, Bao-Liang Lu, Wanzeng Kong
Summary: This paper proposes a Joint label-Common and label-Specific Features Exploration (JCSFE) model for semi-supervised cross-session EEG emotion recognition. The experimental results demonstrate that JCSFE achieves superior emotion recognition performance and provides a quantitative method to identify the label-common and label-specific EEG features in emotion recognition.
IEEE TRANSACTIONS ON NEURAL SYSTEMS AND REHABILITATION ENGINEERING
(2023)
Article
Automation & Control Systems
Yong Peng, Honggang Liu, Wanzeng Kong, Feiping Nie, Bao-Liang Lu, Andrzej Cichocki
Summary: In this article, a joint EEG feature transfer and semi-supervised cross-subject emotion recognition model is proposed to enhance emotion recognition performance by optimizing the shared subspace projection matrix and target label. The spatial-frequency activation patterns of critical EEG frequency bands and brain regions in cross-subject emotion expression are quantitatively identified by analyzing the learned shared subspace.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2023)
Article
Engineering, Biomedical
Dan Peng, Wei-Long Zheng, Luyu Liu, Wei-Bang Jiang, Ziyi Li, Yong Lu, Bao-Liang Lu
Summary: This paper explores the sex differences in emotional EEG patterns and finds that these differences exist in various types of emotions and different cultures. Females have more stable emotional patterns compared to males, and males exhibit contrasting patterns for happiness, sadness, fear, and disgust. The key features for emotion recognition are located in the frontal and temporal sites for females and more evenly distributed throughout the brain for males.
JOURNAL OF NEURAL ENGINEERING
(2023)
Article
Engineering, Electrical & Electronic
Dongrui Wu, Bao-Liang Lu, Bin Hu, Zhigang Zeng
Summary: A brain-computer interface (BCI) allows direct communication between a user and a computer through the central nervous system. An affective BCI (aBCI) monitors and regulates the emotional state of the brain, which has various applications in human cognition, communication, decision-making, and health. This tutorial provides a comprehensive and up-to-date guide on aBCIs, covering basic concepts, components of a closed-loop aBCI system, representative applications, and challenges and opportunities in aBCI research and applications.
PROCEEDINGS OF THE IEEE
(2023)
Article
Computer Science, Artificial Intelligence
Yong Peng, Keding Chen, Feiping Nie, Bao-Liang Lu, Wanzeng Kong
Summary: This study proposes a novel Fuzzy k-means (FKM) method called two-dimensional embedded fuzzy data clustering (2DEFC), which retains structural information and optimizes the projection matrices of two subspaces collaboratively. By optimizing the input and clustering processes for 2D data, competitive performance in 2D data clustering is achieved.
IEEE TRANSACTIONS ON EMERGING TOPICS IN COMPUTATIONAL INTELLIGENCE
(2023)
Article
Computer Science, Artificial Intelligence
Yong Peng, Wenna Huang, Wanzeng Kong, Feiping Nie, Bao-Liang Lu
Summary: Most existing graph-based clustering models adopt a two-stage strategy for clustering, which first completes spectral embedding from a fixed graph and then uses other clustering methods such as k-means to obtain discrete cluster results. However, this discretization operation often leads to deviations from the true solution and the fixed graph is usually suboptimal. Additionally, clustering in separate steps breaks the underlying connections among graph construction, spectral embedding, and discretization. To address these issues, we propose JGSED, a new spectral clustering model that integrates graph construction, spectral embedding, and spectral rotation into a unified objective. JGSED is an end-to-end framework that directly takes data as input and outputs the final binary cluster indicator matrix. An efficient algorithm is proposed to optimize the model variables in JGSED, leading to improved performance compared to other state-of-the-art spectral clustering models.
IEEE TRANSACTIONS ON EMERGING TOPICS IN COMPUTATIONAL INTELLIGENCE
(2023)
Article
Computer Science, Artificial Intelligence
Yong Peng, Wenjuan Wang, Wanzeng Kong, Feiping Nie, Bao-Liang Lu, Andrzej Cichocki
Summary: In this paper, a joint feature adaptation and graph adaptive label propagation model (JAGP) is proposed for cross-subject emotion recognition from EEG signals. By unifying the previously scattered feature learning, emotional state estimation, and optimal graph learning into a single objective, the recognition performance is greatly improved, and the critical frequency bands and brain regions can be automatically identified.
IEEE TRANSACTIONS ON AFFECTIVE COMPUTING
(2022)
Article
Engineering, Electrical & Electronic
Yikai Zhang, Ruiqi Guo, Yong Peng, Wanzeng Kong, Feiping Nie, Bao-Liang Lu
Summary: This study proposes a new model AWIRVFL for EEG-based driving fatigue detection, which addresses the limitations of existing methods. The AWIRVFL model incorporates an auto-weighting variable to consider the importance of different feature dimensions. Experimental results demonstrate that AWIRVFL outperforms existing techniques in driving fatigue detection.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2022)
Article
Engineering, Biomedical
Yong Peng, Fengzhe Jin, Wanzeng Kong, Feiping Nie, Bao-Liang Lu, Andrzej Cichocki
Summary: This paper proposes an OptimalGraph coupled Semi-Supervised Learning (OGSSL) model for EEG emotion recognition, which unifies adaptive graph learning and emotion recognition into a single objective. By optimizing graph learning and improving the label indicator matrix of unlabeled samples, it achieves high accuracy in emotion recognition and automatically identifies EEG frequency bands and brain regions correlated with emotions.
IEEE TRANSACTIONS ON NEURAL SYSTEMS AND REHABILITATION ENGINEERING
(2022)
Article
Engineering, Electrical & Electronic
Yong Peng, Yikai Zhang, Wanzeng Kong, Feiping Nie, Bao-Liang Lu, Andrzej Cichocki
Summary: This paper proposes a semi-supervised sparse low-rank regression model called (SLRR)-L-3 to unify discriminative subspace identification and semi-supervised emotion recognition. Experimental results show that the emotion recognition performance is greatly improved by the joint learning mechanism of (SLRR)-L-3, and the model exhibits additional abilities in affective activation patterns exploration and EEG feature selection.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2022)
Article
Social Sciences, Interdisciplinary
Fangyao Shen, Yong Peng, Guojun Dai, Baoliang Lu, Wanzeng Kong
Summary: This paper introduces a new model for emotion recognition that effectively addresses the impact of distribution discrepancies between different sessions on performance. Experimental results demonstrate that the model performs well in accuracy and data alignment, and can automatically identify critical EEG frequency bands and channels for emotion recognition.