Article
Computer Science, Artificial Intelligence
Azar Rafie, Parham Moradi, Abdulbaghi Ghaderzadeh
Summary: Multi-label classification methods assign more than one label to each instance. High dimensional problems can reduce the performance of machine learning methods, so feature selection is introduced to choose a small set of prominent features. Traditional multi-label feature selection methods fail when applied on data streams, so online methods are introduced. This paper proposes a multi-objective search strategy using mutual information and Pareto optimal set theories to select streaming features, with the obtained results demonstrating its effectiveness.
EXPERT SYSTEMS WITH APPLICATIONS
(2023)
Article
Computer Science, Information Systems
Wenbin Qian, Qianzhi Ye, Yihui Li, Jintao Huang, Shiming Dai
Summary: This paper proposes a relevance-based label distribution feature selection method, aiming to reduce model complexity and computational cost, as well as enhance the generalization ability of the learning model. By considering the relevance between features and labels, a convex optimization function is used for feature selection, and the Pearson correlation coefficient is used to describe the correlation information between labels. Experimental results demonstrate the effectiveness of the proposed method compared to existing methods in multiple evaluation metrics.
INFORMATION SCIENCES
(2022)
Article
Computer Science, Hardware & Architecture
Sadegh Eskandari
Summary: Multi-label learning deals with data in which an instance may belong to multiple class labels simultaneously. The main problem of the existing multi-label feature selection algorithms is their inability to consider all possible subsets of feature space. This paper proposes a new method to address higher-order relevance and redundancy analysis and experimental results demonstrate its superiority over seven state-of-the-art multi-label feature selection algorithms.
JOURNAL OF SUPERCOMPUTING
(2022)
Article
Computer Science, Interdisciplinary Applications
Zana Azeez Kakarash, Farhad Mardukhia, Parham Moradi
Summary: This paper proposes a method of filtering and multi-label feature selection to address the issue of reduced machine learning performance in high-dimensional data. The method utilizes a graph-based density peaks clustering to group similar features and uses ant colony optimization search process to rank features based on their relevancy and redundancy. Experimental results show the superiority of the proposed method over baseline and state-of-the-art methods.
JOURNAL OF COMPUTATIONAL DESIGN AND ENGINEERING
(2023)
Article
Computer Science, Artificial Intelligence
Jinghua Liu, Songwei Yang, Yaojin Lin, Chenxi Wang, Cheng Wang, Jixiang Du
Summary: Multi-label feature selection is an important task in processing multi-semantic high-dimensional data. This paper proposes a new algorithm called FSEP, which considers label significance and pairwise label correlations. The proposed method achieves encouraging results compared with state-of-the-art MFS algorithms, as demonstrated by extensive experiments.
INTERNATIONAL JOURNAL OF MACHINE LEARNING AND CYBERNETICS
(2023)
Article
Computer Science, Information Systems
Wanfu Gao, Hanlin Pan
Summary: Multi-label feature selection is important for addressing multi-label data with high-dimensional features. Existing methods calculate label correlations differently, resulting in different label importance. Our proposed method overcomes the limitations of previous schemes by using mutual information to capture different cores of label sets instead of calculating label importance. Experimental results show that our method achieves the best classification performance among all multi-label feature selection methods.
Article
Computer Science, Artificial Intelligence
Fei Han, Tianyi Wang, Qinghua Ling
Summary: In this paper, an improved feature selection method based on angle-guided MOPSO and feature-label mutual information is proposed, which considers the prior information in feature data and improves the interpretability of selected features. Experimental results show that the proposed algorithm achieves satisfactory results in terms of both improving classification accuracy and reducing the number of selected features.
APPLIED INTELLIGENCE
(2023)
Article
Physics, Multidisciplinary
Jinghua Liu, Songwei Yang, Hongbo Zhang, Zhenzhen Sun, Jixiang Du
Summary: Multi-label streaming feature selection is a popular research area due to its alignment with practical application needs. Existing methods either assume label independence or struggle to understand the relationship between label correlations and feature association. This study introduces an online streaming feature selection method based on label group correlation and feature interaction (OSLGC), which divides labels into groups using graph theory, quantifies feature relationships under different label groups, and designs a feature selection framework that considers online relevance and interaction analysis. Experimental results on ten datasets demonstrate the superiority of OSLGC over other mature multi-label streaming feature selection algorithms in terms of predictive performance, statistical analysis, stability, and ablation experiments.
Article
Computer Science, Artificial Intelligence
Dianlong You, Yang Wang, Jiawei Xiao, Yaojin Lin, Maosheng Pan, Zhen Chen, Limin Shen, Xindong Wu
Summary: This paper proposes a novel online multi-label streaming feature selection scheme, taking into account the correlation between labels, and achieves significant advantages in terms of effectiveness and efficiency in performance evaluations.
IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING
(2023)
Article
Computer Science, Artificial Intelligence
Yuling Fan, Baihua Chen, Weiqin Huang, Jinghua Liu, Wei Weng, Weiyao Lan
Summary: The task of multi-label feature selection (MLFS) is to reduce redundant information and generate the optimal feature subset from the original multi-label data. Existing methods either explore label correlations using pseudo label matrix or consider feature redundancy using information theory technique, but no prior literature unites them in a framework to perform feature selection. Therefore, a novel MLFS method called LFFS is proposed based on label correlations and feature redundancy. Experimental results demonstrate the effectiveness and superiority of the LFFS method among ten competition methods.
KNOWLEDGE-BASED SYSTEMS
(2022)
Article
Computer Science, Artificial Intelligence
Ping Zhang, Guixia Liu, Wanfu Gao, Jiazhi Song
Summary: This study introduces two new multi-label feature selection methods, LSMFS and MLSMFS, which extract all supplementary information and the maximum supplementary information from other labels, respectively. These methods aim to address the different effects and dynamic changes of label relationships, and experiments demonstrate their effectiveness against nine other methods.
PATTERN RECOGNITION
(2021)
Article
Computer Science, Information Systems
Nitin Kumar Mishra, Aditya Singh, Pramod Kumar Singh
Summary: Understanding personality is beneficial for various purposes, such as predicting users' personality and text classification. However, existing methods have low accuracy, prompting the proposal of PTLFM, a multi-label classification method based on language and feature selection. Experimental results show that PTLFM achieves significant improvement in personality trait recognition and outperforms existing machine learning and deep learning models.
MULTIMEDIA TOOLS AND APPLICATIONS
(2022)
Article
Computer Science, Artificial Intelligence
Dipanjyoti Paul, Anushree Jain, Sriparna Saha, Jimson Mathew
Summary: This paper presents an adaptive feature selection algorithm for multi-label classification scenarios, real-time selecting the optimal feature subset online. Through a three-phase filtering process, the algorithm improves the accuracy and efficiency of feature selection.
KNOWLEDGE-BASED SYSTEMS
(2021)
Article
Computer Science, Artificial Intelligence
Minlan Pan, Zhanquan Sun, Chaoli Wang, Gaoyu Cao
Summary: This paper discusses the challenges of high-dimensional multi-label data and the popularity of filtering feature selection algorithms in pattern recognition and machine learning. It introduces the research topic of multi-label feature selection based on mutual information and proposes an approximation method that considers label correlation without significantly increasing computational cost.
INTELLIGENT DATA ANALYSIS
(2022)
Article
Computer Science, Artificial Intelligence
Utkarsh Agrawal, Vasudha Rohatgi, Rahul Katarya
Summary: The problem of feature selection involves selecting the most informative subset of features that have the most impact in classification. This paper proposes a novel variant of the Equilibrium Optimizer called Normalized Mutual Information-based equilibrium optimizer (NMIEO) for feature selection. The proposed method incorporates a local search strategy based on Normalized Mutual Information and utilizes Chaotic maps for population initialization. Experimental results demonstrate the superior performance of NMIEO compared to other competitive methods.
EXPERT SYSTEMS WITH APPLICATIONS
(2022)
Article
Computer Science, Information Systems
Jaesung Lee, Jonghoon Chae, Dae-Won Kim
MULTIMEDIA TOOLS AND APPLICATIONS
(2017)
Article
Computer Science, Artificial Intelligence
Jaesung Lee, Dae-Won Kim
PATTERN RECOGNITION
(2017)
Article
Mathematics, Interdisciplinary Applications
Jaesung Lee, Wangduk Seo, Dae-Won Kim
Article
Engineering, Electrical & Electronic
J. Lee, W. Seo, D. -W. Kim
ELECTRONICS LETTERS
(2018)
Article
Computer Science, Information Systems
Jaesung Lee, Wangduk Seo, Jin-Hyeong Park, Dae-Won Kim
MULTIMEDIA TOOLS AND APPLICATIONS
(2019)
Article
Computer Science, Information Systems
Xu Miao, Jaesung Lee, Bo-Yeong Kang
Article
Mathematics, Interdisciplinary Applications
Jaesung Lee, Dae-Won Kim
Article
Engineering, Electrical & Electronic
Jaesung Lee, Wangduk Seo, Ho Han, Dae-Won Kim
JOURNAL OF SENSORS
(2018)
Article
Computer Science, Software Engineering
Jinhyuk Jo, Song Ko, Dae-Won Kim, Jaesung Lee
Summary: Computer-aided pronunciation training systems are considered useful for beginner-level students who may feel uncomfortable with face-to-face training or having their pronunciation corrected in front of others. This study proposes a system targeting unacceptable pronunciation caused by confusion among contextual allophones, resulting in improved pronunciation skills according to experimental results.
CONCURRENCY AND COMPUTATION-PRACTICE & EXPERIENCE
(2021)
Article
Physics, Multidisciplinary
Jaesung Lee, Jaegyun Park, Hae-Cheon Kim, Dae-Won Kim
Article
Computer Science, Information Systems
Jaesung Lee, Injun Yu, Jaegyun Park, Dae-Won Kim
INFORMATION SCIENCES
(2019)
Article
Physics, Multidisciplinary
Jaegyun Park, Min-Woo Park, Dae-Won Kim, Jaesung Lee
Article
Computer Science, Artificial Intelligence
Hae-Cheon Kim, Jin-Hyeong Park, Dae-Won Kim, Jaesung Lee
PATTERN RECOGNITION LETTERS
(2020)
Article
Computer Science, Information Systems
Wangduk Seo, Dae-Won Kim, Jaesung Lee